Home

from isi.edu - Information Sciences Institute

1. AT Bet proposition 00e cesses owes AT get propositions Ir ere e vue 47 get propositions in module AT get propositions of lese E Rd a 47 FA riggs A ee 4T get relationi i i 9 gea de ER 4T 120 get relation extension s 4T get rul s ii edveeekeriteiervWer iia ids 34 47 get self or prototype cesses 120 get short name cese geb ra ede agate 131 get slot maximum cardinality 120 get slot minimum cardinality 120 get slot value iiis eed eb nes 120 get slot value type o rr nn 120 get symbol fiche kIxxg uo e eRERRER GE REP 131 BOt types iine e 1eeezec abe RE ILPRITRUEA 48 get why justification 120 get whynot justifications 120 G ES FALSE DEMON e ee e rhe 73 GOES TRUE DEMON 2 ccccecianie aeadaes prever 73 GOES UNKNOWN DEMON 00 00005 73 H has forward justifications 120 help e eb R LEEPERRMI te RESI E SEMI 34 help print outline on DESCRIPTION 121 help print outline on NAMED DESCRIPTION si edef erereetRRRia dee ueri RERO apre 120 HOLDS i csr bb hh ee ob aaa aud 73 I IMAGE UR asec tte ted ae federgwefateurateradbinn 74 In dialeCt wide cise dee ecu ru er Eg 121 in module evel See ear tie here eet 35 inconsistent truth value 121 inconsistent 8 v 9 evi eee e 121 INEQUALITY isser ne RR RR RERESSRPR 74 initial
2. 127 specializesT c cler See beRePIREUE EE 127 SQUARE RDOT e cetMliwis edad ieee eee 80 STANDARD DEVIATION 00 0000 81 Start onRtosauruS m e ivre ES genah 127 start powerloom gui sssss s 128 start powerloom server eeeeeee 128 stop ontosasa r s 42 ls 9 nna ida 128 stop powerloom gui een 128 stop powerloom server ce ea 128 strengthen truth value s 128 Strict truth value ii e eer RE 128 STRING CUNGATENATE ee R1 Herne 81 STRING MATOH e peek add e a oi MS 81 STRING MATCH IGNORE CASE 0 81 string to objeCt sodass he Re Ris 58 SUBRELBATION 2cccci ote soeek deen deka eae RR 81 SUBSET DE n le6k3 ai Aebe Xnsssitid eeded tel de 81 SUBSIRING een ee t Ree beri Sed eae 81 SUBSUMPTTON QUERY ccc n rm RR cae sets oda tis 81 SUM P 82 SUPERRELAT TON 2252 anrai EE E EANES 82 SYMMETRI Cresi deetan eeen o E Neea So 82 SYNONYM as cue rn feed cea aoe doe eo A EEE ERES 82 T t rmifyio ied e 4er sere Idae Pee 128 test closed slot 0 cece eee 128 test environment level 133 146 test function slot 000 128 test marker on MARKER TABLE 128 test relation on arguments 128 te est slot value 7 x aia raa eR ER 129 test special marker table 129 test su brelation 4 cc ccs sr see e s
3. 41 assert from query sex ene rU S eb APUL 28 assert nary proposition 41 assert propositio n i l5ec en DIL r9 41 assert rul64 cc re Rinne eas ed ewes 29 assert unary proposition 41 ASSERTION QUERY ie 0csuelevevederasnasvonces 68 B BACKTRACKING QUERY eese 68 BINARY RELATION Rb eher So eo eer 69 bottom i i6sbenss9eepeexer S EGR 112 BDUND VARIABLES eis0 cscs cGeuteeds sede ede ede 69 C call all facts of 22 9 29 9 px 112 Gall aSk i reu eb RR eR Rr OPENED RERO 113 call defconcept i e et eg 113 call deff nction i aerem ner n 113 call defobj Ct i sislsase kac e e ae ees 113 call defproposition lese e re e ER 113 call defrelation i o e ca dae en 113 call list undefined relations 113 call propagate constraints 113 call retrieve ba es Bs Rebun sas 113 call retrieve parti l lesa is 113 call run forward rules ssssss 113 call set inference level 114 CARDINALITY em cR abeiem dared see dete 69 Function Index CS EE 29 change module 2 4 BRA SERC RP SES DER 41 CLASS wus Sau nene eE IES ERE Ne NES 114 Classify iNStanGes oo eek e eens ess 30 classify relations d bie ee b iieri 29 clear cach68 a eee e e bes e tbe 30 41 clear instances sees 30 Clear modile sisi ee ex sed eee a 30 41 CLOSED i si succs ye mesei enaa aud aye arene cavay aac
4. 50 retract proposition 9 er hr e 50 retract rul6e sess ism e Ue Re ppERSRR E 36 retract unary proposition 50 Z6brie6VO sse twee ae crams Hered eee dae 37 51 retrieve partial c d n een 126 run forward rules 0000 51 126 run powerloom testsS eee eee ees 126 S S aBSE nee ee ee ee 51 S assert proposition i ne ke eher 52 S change mModul ics cee gestae he deme ees 52 s clear module 1 d d gw sen Reg 52 S CONCEIVE is eroe pr GG e a Ue RORIS ERE 52 S create Concept s ggc s d e eprY mithi 52 S create function i e Rb 53 S create modul6 n iR eR RERS 53 create objJeCt sisse bRe oben baeo 53 s create relation 2 cns 9 rt hen 53 amp destroy object ic os Lee eee es ti 54 S evaluate issdic uod i ori edit p tei 54 S ESLt arlby 2690s Ires ea e enint 54 S amp get child modul S i vas er nasa dee 54 S get CoOnCept a dne e Rex eR era 54 s get concept instances riita 54 s get direct concept instances 54 S get domailm iiesisscc o eh eer a des 55 s get inferred binary proposition values Sadie EEEIEE EEA REE NERE A DEERAS 55 get nth domain sis yino iiai 55 get object siis anr ese beer dee ag 55 s get parent modules LL 55 S get propositiOn c nc gr rns 55 S get propoBitiOns x rissen edaria 56 s get propositions of eee 56 SH LSC AVON SS c
5. 0 0 2 e ee eee 78 RANGE CARDINALITY LOWER BOUND 78 RANGE CARDINALITY UPPER BOUND 79 RANGE MAX CARDINALITY 008 79 RANGE MIN CARDINALITY 008 79 RANGE TY PEs eR I eee Ide RET abe se dada dues 79 recall marked objects on MARKER TABLE 124 reco rd justifications le clem 124 REFLEXIVE o Lr Ee ID bee nEPRRPUeMS 79 REFUTATIUN QUERY sada iden cevdoeaases 79 register computation function 125 132 register computation function name 125 register logic dialect print function 125 register specialist function 125 132 register specialist function name 125 RELATION COMPUTATION 2 2 2 0 ce eee e 79 RELATION CONSTRAINT ciem Rs nee ne 79 RELATION EVALUATOR eese 80 relation n ame ci ls loc tr s 125 RELATION SPECIALIST eseseeeseese 80 relatjon i n tek de pb Rada aA 125 remove at on QUERY SOLUTION TABLE 125 remove deleted members on PAGING INDEX 125 repropagate constraints sess 36 reset featureS cd gn ele es NE RAR 36 reset powerloom ccc eee eee 36 50 145 reset query caches cece eee eee 126 TET TAC carura e m oE EIk nin NER RERR 36 50 retract binary proposition 50 retract facts Of gach ved handsets decease 36 retract facts of instance 126 retract from query db rai binini saaa 36 retract nary proposition
6. retrieve age Joe x There is 1 solution so far 1 X 10 So if one needs to model named individuals where equality might be asserted e g to model a person with an alias one has to resort to using function terms For example deffunction individual name i assert age individual A 12 assert individual A individual B retrieve age individual B a There is 1 solution so far 1 A 12 3 14 Classification Subsumption Chapter 3 Annotated Example 24 3 15 3 16 3 17 3 18 3 19 3 20 3 21 3 22 3 23 Truth Maintenance Inference Control Keyword Axioms Cardinality Type Reasoning with Frame Predicates Loom to PowerLoom Deviations from KIF Differences from Loom Defaults Sets Lists SETOFALL KAPPA Chapter 4 Communicating with PowerLoom 25 4 Communicating with PowerLoom There are a variety of modes users can choose from for interacting with the PowerLoom system The simplest is to use the PowerLoom command interpreter The interpreter supports a type in window that allows line at a time entry of commands You can use the interpreter to load files of PowerLoom declarations to create and edit knowledge base objects to ask queries and to modify settings in the execution environment A second mode of interaction involves writing an application that makes calls to the PowerLoom API see Chapter 6 PowerLoom API page 40 PowerLoom implements an extensi
7. gt company name c name PowerLoom defines a subconcept subrelation relationship between a pair of concepts or a pair of relations by asserting an implication relation between them The above implication expands into the assertion for all values of c and name if the fictitious business name relation holds for c and name then the company name relation also holds for c and name This is equivalent to the assertion forall c name gt fictitious business name c name company name c name Since implication relationships occur very commonly PowerLoom provides several syntactic shortcuts for defining them We have seen one such shortcut earlier our definition of corporation included the clause c company which specified that corporation is a subconcept of company In our definition of fictitious business name the keyword gt introduces a similar shortcut which tells us that fictitious business name is a subrelation of company name Let us assert a fictious business name for MegaSoft assert fictitious business name megasoft MegaSoft If we query for the company names of MegaSoft we get two names one of them asserted directly and one of them infered by the subrelation rule retrieve all company name megasoft x There are 2 solutions 1 X MegaSoft Inc 2 X MegaSoft 3 5 Functions This illustrates another point A PowerLoom relation is by default multi valued which in
8. 12 geographic location Now we can assert some facts about U S geography including one deliberate mistake assert and country united states geographic location eastern us contains united states eastern us state georgia contains eastern us georgia city atlanta contains georgia atlanta geographic location southern us contains united states southern us state texas contains eastern us texas city dallas contains texas dallas city austin contains texas austin We would like to repair the incorrect assertion contains eastern us texas The PowerLoom command retract allows us to erase assertions that should not be true ask contains eastern us texas TRUE retract contains eastern us texas assert contains southern us texas ask contains eastern us texas UNKNOWN Chapter 3 Annotated Example 18 Retraction should not be confused with assertion of negative propositions For example asserting that Texas is not a state would not retract the assertion that it is astate Instead an evident logical contradiction is detected as a clash and the clashing proposition is disallowed assert not state texas gt Derived both TRUE and FALSE for the proposition P STATE TEXAS Clash occurred in module MDL PL KERNEL KB business ask not state texas gt UNKNOWN 3 9 Clipping of Values Programmers are accustomed to changing the values of att
9. Return the predicate operator op of p QUERY prop PROPOSITION options THING Relation Search control relation that allows one to prove or retrieve bindings for prop with modified search control options The list of accepted options is currently the same as are legal for a top level ask or retrieve query The special option value IN HERIT inherits the option value from the parent or top level query The option pair INHERIT ALL inherits all parent options which can then be further modified by additional individual option specifications At most how many solutions will be gen erated is controlled by the HOW MANY option just like in the top level retrieve The default is 1 which is again the same as for retrieve but different from how nor mal subgoals behave those behave in a lazy all solutions mode The reason for this is that for partial match subqueries solutions need to be generated eagerly therefore a default of generating all solutions is not desirable RANGE r RELATION rng COLLECTION Relation True if for any tuple T that satifies r the last argument of T necessarily belongs to the concept rng range exists for convenience only and is defined in terms of nth domain range assertions should be avoided since they create redundant nth domain propositions use nth domain directly RANGE CARDINALITY r RELATION i THING gt card INTEGER Function Function that returns the cardinality of the
10. gt gt lt sentence gt lt gt lt sentence gt lt gt gt lt sentence gt lt lt gt lt sentence gt lt lt gt gt lt sentence gt rc axioms lt sentence gt lt sentence gt lt keyword option gt Function parameters can be typed or untyped If the gt option is supplied it specifies the output variable of the function Otherwise the last variable in the parameter list is used as the output variable See defrelation for a description of lt keyword option gt s definstance amp rest args PARSE TREE LOGIC OBJECT N Command Define or redefine a logic instance definstance is an alias for defobject which see defmodule name NAME amp rest options OBJECT N Command Define or redefine a module named name The accepted syntax is defmodule lt module name gt documentation lt docstring gt includes lt module name gt lt module name gt uses lt module name gt lt module name gt lisp package lt package name string gt java package lt package specification string gt cpp namespace lt namespace name string gt java catchall class api TRUE FALSE case sensitive TRUE FALSE shadow symbol java catchall class lt class name string gt lt other options gt name can be a string or a symbol Modules include objects from other modules via two separate mechanisms 1 they inherit from t
11. lt has age person age and has ssn person ssn exists 1 72 73 74 ist Medical Kb patient record ssn 71 72 age 73 74 Chapter 7 Built In Relations 75 LENGTH x THING gt z INTEGER Function Function that returns the length of a string or a logical list NOT YET IMPLE MENTED FOR LISTS LENGTH OF LIST COLLECTION length INTEGER Relation Computes the length of the list or set l LEXEME r THING s STRING Relation s is a lexeme for the relation or individual r A relation or individual r can have zero or more lexemes words that are natural langage equivalents of a logical constant The same lexeme may be attached to more than one constant LIST SEQUENCE Class Not documented any value OBJECT Class Parameter of Not documented the cons list CONS OF LIKE ANY VALUE SELF Slot of Not documented LIST CONCATENATE x COLLECTION y COLLECTION gt r LIST Function Concatenate lists x and y into r If x and or y are not lists but sets or more general collections the order of the elements in the result list r will be arbitrary LISTOF m THING gt c LIST Function Term forming function that defines an ordered list consisting of all function argu ments Within logical expressions listof is most commonly used in conjunction with the member of predicate For example the query retrieve x member of x listof a b c returns the constan
12. method stringified source STRING Slot of Not documented method code METHOD CODE Slot of Not documented function code FUNCTION CODE Slot of Not documented method function BOOLEAN Slot of Not documented GOES FALSE DEMON r RELATION Relation computation COMPUTED PROCEDURE Names a computation a function that is attached logically to r Each time a proposition with predicate r becomes false the function is applied to that proposi tion GOES TRUE DEMON r RELATION Relation computation COMPUTED PROCEDURE Names a computation a function that is attached logically to r Each time a proposition with predicate r becomes true the function is applied to that proposition GOES UNKNOWN DEMON r RELATION Relation computation COMPUTED PROCEDURE Names a computation a function that is attached logically to r Each time a proposition with predicate r becomes unknown the function is applied to that propo sition HOLDS relation RELATION arguments THING Relation True if the tuple arguments is a member of the relation relation holds is a variable arity predicate that takes a relation as its first argument and zero or more additional arguments It returns values equivalent to a subgoal that has the first argument as a predicate and the remaining arguments shifted one place to the left For holds to succeed the first relation argument must be bound PowerLoom will
13. JUSTIFICATION Function Returns the current WHY justification May also throw one of the following subtypes of EXPLAIN EXCEPTION EXPLAIN NO QUERY EXCEPTION EXPLAIN NO SOLUTION EXCEPTION EXPLAIN NO MORE SOLUTIONS EXCEPTION EXPLAIN NOT ENABLED EXCEPTION EXPLAIN NO SUCH LABEL EXCEPTION EXPLAIN QUERY TRUE EXCEPTION get whynot justifications query QUERY ITERATOR label STRING Function mapping EXPLANATION MAPPING LIST OF JUSTIFICATION Programmer s interface to WHYNOT function Derive justifications why query failed or if label was supplied as non NULL lookup its justification relative to mapping and return the result has forward justifications proposition PROPOSITION Function BOOLEAN Return TRUE if proposition has any forward justifications help print outline top Method on NAMED DESCRIPTION stream OUTPUT STREAM current depth INTEGER depth INTEGER named BOOLEAN Helper function for print concept outline Chapter 11 Miscellaneous 121 help print outline top stream OUTPUT STREAM Method on DESCRIPTION current depth INTEGER depth INTEGER named BOOLEAN Helper function for print concept outline in dialect dialect NAME KEYWORD N Command Change the current logic dialect to dialect Currently supported dialects are KIF STELLA and PREFIX STELLA The STELLA dialects are not recommended for the construction of knowledge bases they are mainly used internally by PowerLoom inconsiste
14. K known truth value ese eens 121 L LENGE tn posae RAS RIDERE EUR 75 length on PL ITERATOR eee eee eeee 49 length on QUERY SOLUTION TABLE 121 LENGTH OF LIST oarre 2b ue RR ER Rs 75 EREMB Ys ce iu bed PARES RI PASS 15 LIST CUNCATENATE 1 2 ERI ad aoasieacd T9 list f ature s essenin inae REINO RE ed 121 listm d les eissor eiside ERI RE EIEEE 35 list to pl iteratOr ees 49 list unclassified instances 121 list unclassified relations T21 list undefined relations 121 listify on PL ITERATOR eee 133 listify on QUERY ITERATOR cesses 122 TET uL PME 75 Loa8d eres ER Rises ota EE EIER UATP QE 35 49 load cmd line files sss 122 load direCtory lik Rb t ERES 122 load fil e 2 2 e rise had ERES RP A 35 load incmodi le ieLbbre4 iex es 131 load native stream esses 49 load native stream in module 131 load stre am annosa nan hes 49 122 load stream in module 122 131 logic Cl 8s8 Sas pec et EUDEN TA 122 144 logic form lessS 7 2 o ebria 122 logicemodule re b rd bine 122 lookup on QUERY SOLUTION TABLE 122 lookup native computation 122 lookup native specialist 122 M MAIN eid oh caton hier eer age haan e rbi dis 132 MAXIMUM ELEMENT m ter IRI e e eia RR RREPY 75 MAXIMUM VALUE
15. NAMED DESCRIPTION N Command Define or redefine a concept The accepted syntax is defconcept lt conceptconst gt lt var gt lt parent gt documentation lt string gt lt lt sentence gt gt lt sentence gt lt lt lt sentence gt gt gt lt sentence gt lt gt lt sentence gt lt gt gt lt sentence gt lt lt gt lt sentence gt lt lt gt gt lt sentence gt axioms lt sentence gt lt sentence gt lt keyword option gt Chapter 5 Commands 31 Declaration of a concept variable lt var gt is optional unless any implication arrow options are supplied that need to reference it A possibly empty list of concept names following lt var gt is taken as the list of parents of lt conceptconst gt Alternatively parents can be specified via the gt option If no parents are specified the parent of lt conceptconst gt is taken to be THING lt keyword option gt represents a keyword followed by a value that states an assertion about lt conceptconst gt See defrelation for a description of lt keyword option gt s deffunction amp rest args PARSE TREE NAMED DESCRIPTION N Command Define or redefine a logic function The accepted syntax is deffunction lt funconst gt lt vardecl gt gt lt vardecl gt documentation lt string gt lt lt sentence gt gt lt sentence gt lt lt lt sentence gt
16. gt z retrieve x x foo a b However if we assert that foo is total then we get a skolem back when we execute the same retrieve assert total foo retrieve x x foo a b TRANSITIVE r RELATION Relation A binary relation r is transitive if r x y and r y z implies that r x z Note that functions cannot be transitive since their single valuedness would not allow multiple different values such as r x y and r x z due to the Unique Names Assumption made by PowerLoom TYPE OF c COLLECTION x THING Relation True if x is a member of the concept c VALUE function FUNCTION arguments THING gt value THING Function True if applying function to arguments yields the value value The value predi cate is the analog of holds except that it applies to functions instead of relations VARIABLE ARITY r RELATION Relation Asserts that the relation r can take a variable number of arguments Chapter 7 Built In Relations VARIANCE COLLECTION variance NUMBER Binds variance to the variance of the numbers in 71 83 Relation Chapter 8 PowerLoom GUI 84 8 PowerLoom GUI The PowerLoom GUI or knowledge editor is a Java based graphical client for PowerLoom The GUI is implemented as a Swing based Java application which communicates with an embedded or remote PowerLoom server using a SOAP communication layer over HTTP The architecture of the Pow
17. ut bees tL XRPYSCHSTIE 75 MEAN VALU Bio 2 cS oes ace ese totes beer EI METERS 75 MEDTAN VALUE ewe eIe rinra DPI 75 MEMBER OPF nere RR tIemR Pte soee teeter ees 75 MINIMUM ELEMENT x Rb RS eT gins 76 MINIMUM VALUE ssesispeSRORS eI me Re iet ersa 76 MUTUALLY DISJOINT COLLECTION 76 N NAME TO OBJEGE cep eE EA N iad 76 Named description wince taupes eels 123 natural deduction mode 123 next on PL ITERATOR 0 cece nnne 50 non empty on FLOAT VECTOR sss sese 123 non empty on QUERY SOLUTION TABLE 123 NORMAL QUERY sic cogs dues ach Seed casein 76 nth on QUERY SOLUTION TABLE sess 123 NTH D MATN e ib e ern chad Seas bead Gh 76 NIH SELEMENT otrai chee cGees ge RI ee E Eun 76 NTH HEAD eorum ed nU SORTE bud 76 NTH REST esecrdcinc cies eue T cat eee Ree wane fuus 76 NUMERIG MAXIMUM 2er p ere a segre tees 76 NUMERIG MINIMUM eee 00 ones ceeeaaer eae 76 NUMERIGHSED orrera irani e seme ead oe re O object name cis ed aa e awe EMPIRE ad gt 123 OBJECT NAME incc ced aya mh epe tins inen due toe T object name String l ice re Ra Eee 123 bject surrog t s iss oer nee 123 object surrogate setter 123 object to float isisiesi dx Rp Pens 50 object to integer c o osdga b e u kaske 50 object to parsable string 50 object to string cod ee eA einer ERR 50 ORDERED ssr i pa tannas mna i ain LEER RE Te P PHRASE ice Le
18. Execute a yes no query composed of input variables inputVariables and body queryBody Before executing bind variables to inputBindings in sequence apply ask inputVariables queryBody inputBindings apply kappa description DESCRIPTION vector VECTOR BOOLEAN Function Apply inherit the description description to members of the vector vector Return TRUE if no clash was detected Constraint propagation happens only if it is enabled prior to calling apply kappa apply retrieve amp body body CONS oBJECT Macro Execute a query composed of io variables variables and body queryBody Before executing bind variables to inputBindings in sequence If one variable is left unbound returns a cons list of bindings of that variable If two or more are unbound returns a cons list of cons lists of bindings Setting the option singletons to FALSE always returns a list of lists Example call apply retrieve variables queryBody inputBindings ask partial amp rest proposition amp options PARSE TREE FLOAT N Command Similar to ask which see but return the highest partial match score for the supplied proposition instead of a truth value If the option MAXIMIZE SCORE is set to FALSE return after the first partial match score has been generated bottom self OBJECT BOOLEAN Function Return TRUE if self is the undefined individual BOTTOM call all facts of instanceRef OBJECT LIST OF PROPOSITION Command Return
19. PowerLoom with g 3 2 and later under Linux Redhat 8 0 amp 9 0 SunOS and MacOS X and with CygWin 5 0 and MinGW 5 0 under Windows 2000 and XP Both CygWin and MinGW provide a GNU Unix environment but MinGW can generate Windows executables that are fully standalone We ve also managed to compile PowerLoom under MS Visual C however we never got the Boehm garbage collector to work The GC claims to be very portable so this should be solvable for somebody with good knowledge of MS Visual C For the Java version you will need Java JDK 1 2 or later To get reasonable performance you should use J2SDK 1 4 or 1 5 We ve run the Java version of PowerLoom on a variety of platforms without any problems Chapter 10 Installation 105 Any one of the Lisp C or Java implementations of PowerLoom can be used to develop your own PowerLoom based applications Which one you choose is primarily a matter of your application and programming environment The Lisp and Java versions are comparable in speed the C version is usually a factor of 2 3 faster than Lisp or Java 10 2 Unpacking the Sources Uncompress and untar the file powerloom X Y Z tar gz or unzip the file powerloom X Y Z zip in the parent directory of where you want to install PowerLoom X Y Z are place holders for the actual version numbers This will create the PowerLoom tree in the directory powerloom X Y Z we will use Unix syntax for pathnames All pathnames
20. all direct superrelations 109 all dir ct types glas RR UR Red 109 all equivalent relations 110 all f cts 0f v dc A E RE ERES 27 28 all f cts of instance iucc iac eR nnus 110 all f cts of n didledc s pi a e EEn 110 all inconsistent propositions 110 all inStanceS i 2e sers em se donor en 110 141 all named descriptions 110 all named instances cesses ss 110 all named term s i e RR EROS 111 all pr positionsS ilibiuelesBR RR GU READERS 111 all relation values sess Tid allcsentences Of o v b cial aeeeecaas 111 all slot value types 0 111 all sl t yvalue8 ie menie Res TTL all subrelations es e e Re AR EEER 111 all superrelations esses TIT all taxonomic types sssesssese 111 allcterms 09 Rx beu MUR REIR M AER DT all types ceienenme ac PEE RPG HER REDE 111 all nnam d t rms 2 9 eI er Reg eds 112 allocate supported closure iterator 112 allocate transitive closure iterator 112 ANTISYMMETRIO Lee e Re eee e RePSHHPS 68 appty aSk e ehpPRR ree ERR RR ERE 112 apply k ppa smoro ee Gide ds 112 apply retrieve ele ee rx Eiei Enie 112 ARIT Y ilc mene eue eR RIS RUDI ES 68 ASK i tile teed e ek ex DAE ER COE UR E e as 28 40 ask partial fico ilb RR 3X A3 RAS TR RES 112 eE E EEEE aes aie EE ave aie ads 28 assert binary proposition
21. child contexts LIST OF CONTEXT Slot of Not documented base module MODULE Slot of Not documented all super contexts CONS OF CONTEXT Slot of Not documented context number INTEGER Slot of Not documented util texinfo insert doc PL KERNEL context of COVERING c COLLECTION cover SET Relation True if c is a subset of the union of all collections in the set cover see disjoint covering CUT arguments THING Relation Prolog like CUT Succeeds the first time and then fails Side effect Locally disables query optimization DIRECT SUBRELATION r RELATION sub RELATION Relation True iff sub is a direct subrelation of r written in set notation sub lt r and there is no s such that sub lt s lt r This relation will generate bindings for at most one unbound argument DIRECT SUPERRELATION r RELATION super RELATION Relation True iff super is a direct superrelation of r in set notation super gt r and there is no s such that super gt s gt r This relation will generate bindings for at most one unbound argument DISJOINT cl COLLECTION c2 COLLECTION Relation True if the intersection of c1 and c2 is empty DISJOINT COVERING c COLLECTION disjointcover SET Relation True if c is covered by the collections in disjointCover and if the member sets in disjointCover are mutually disjoint For example the concepts Igneous Rock Me
22. for possible is a relationships clear caches Command Clear all query and memoization caches clear instances amp rest name NAME N Command Destroy all instances belonging to module name or any of its children Leave meta objects e g concepts and relations alone If no name is supplied the current module will be cleared after confirming with the user clear module amp rest name NAME N Command Destroy all objects belonging to module name or any of its children If no name is supplied the current module will be cleared after confirming with the user Important modules such as STELLA are protected against accidental clearing conceive formula PARSE TREE OBJECT N Command Guess whether formula represents a term or a sentence proposition If we are not sure assume its a proposition If its a term return its internal representation If a proposition construct a proposition for formula without asserting its truth value Return the conceived proposition object KIF example conceive happy Fred builds the proposition expressing that Fred is happy without explictly asserting or denying it Note that for this to succeed the relation happy must already be defined see assert If the logic constant Fred has not yet been created it is automatically created as a side effect of calling conceive copyright Command Print detailed PowerLoom copyright information defconcept amp rest args PARSE TREE
23. lt lt gt gt lt sentence gt axioms lt sentence gt lt sentence gt lt keyword option gt lt gt lt sentence gt lt gt gt lt sentence gt lt lt gt lt sentence gt lt objdefinition gt defobject lt constant gt documentation lt string gt lt keyword option gt lt ruledefinition gt defrule lt constant gt lt sentence gt Chapter 13 PowerLoom Grammar 140 documentation lt string gt lt keyword option gt lt ruledefinition gt defrule lt ruleconst gt lt sentence gt Function Index Function Index rr 68 pL 68 CIC CPP ts ae hee EE PP 68 ee ee ee E E 68 lt E 68 or Er EEUU REM 68 gt Susie pea het ae ect leote eet atm eles 68 DE LII e ee eee eee aes 68 2 2 d element on 2_D_ARRAY 00 0000 109 2 d element on 2 D FLOAT ARRAY 109 2 d element setter on 2 D ARRAY 109 2 d element setter on 2 D FLOAT ARRAY 109 ABSTRACT digiti DER ta Dies 68 dd load path iieict t bp EEE E RO RE 24 add testing example sess 109 add training example 109 AGGREGATE o eie en e anaE EEREN ERRE ERLIE E 68 all asserted types s e RR neat 109 llsclass instance s ei eisian naiera 109 allsCyCl68 eed ea anr stlawa bru el Eai 109 all direct subrelations e nus 109
24. set the level globally class objectRef OBJECT BOOLEAN Function Return TRUE if objectRef denotes a class coerce to instance self OBJECT original OBJECT LOGIC OBJECT Function Return the logic instance referred to by self coerce to instance or literal self OBJECT original OBJECT Function OBJECT Return the logic instance referred to by self or self if it is a literal e g string or number that can t be coerced coerce to vector self OBJECT VECTOR Function Return a vector containing the elements in self Coerce each element of self to be a logic object or literal collection objectRef OBJECT BOOLEAN Function Return TRUE if objectRef denotes a relation or a class conceive term tree OBJECT OBJECT Command tree is a term expression a string or an s expression or is a class reference a symbol or surrogate Return a possibly newly conceived term representing the internalized representation of that term conjoin truth values tvl TRUTH VALUE tv2 TRUTH VALUE Function TRUTH VALUE Return the logical conjunction of truth values tv1 and tv2 consify self CONS Method on JUSTIFICATION Return a CONS tree representation of the proof self Each proof step is represented as a CONS tree of the form lt proposition gt lt key gt lt value gt lt antecedent gt where each antecedent is a CONS tree representing a subproof The consification follows the original
25. since KR failures or awkward models are often due to we could not express X in language L The theoretical undecidability Chapter 1 Introduction 2 and intractability of such an expressive language is counteracted by providing lim ited pragmatic reasoning services that cover commonly encountered situations For example reasoning with second order sentences that quantify over relations is unde cidable and leads to very unfocused search however such sentences are very useful to describe axiom schemata that can be cheaply run in forward direction to create regular first order rules in a process not unlike macro expansion e Reasoning The primary reasoning mechanism is logical deduction which infers state ments that logically follow from the asserted statements and rules Such statements can be asked about using PowerLoom s query commands ask for true false questions and retrieve for Wh questions PowerLoom uses a natural deduction system to answer queries but also has a large number of specialized reasoning procedures to efficiently handle concept and relation hierarchies sets frame predicates search con trol etc The specialist architecture is extensible to allow users to plug in their own reasoners or computed predicates PowerLoom also supports hypothetical reasoning equality reasoning arithmetic and reasoning with inequalities While PowerLoom is not a description logic it does have a classifier that can classify concept a
26. x all x happy x 10 happy x inference level assertion Chapter 6 PowerLoom API 58 10 happy x inference level assertion timeout 1 0 If there is only a single output variable as in all the examples above each element generated by the returned iterator will be a binding for that variable unless the output variable was declared with a surrounding pair of parentheses For example all x happy x In that case the generated elements will be one element lists If there are multiple output variables each element generated by the returned iterator will be a list of variable bindings that can be accessed using the various get nth functions The list of output variables does not need to be declared in which case they are taken to be the open variables in the query proposition in the order in which they were encountered If order does matter or should be different from its default it can be forced by declaring the set of output variables Names in query will be interpreted relative to module module name A null module name or the empty string refers to the current module If no module can be found with the name module name then a STELLA no such context exception is thrown s save module module name STRING filename STRING Function ifexists STRING environment ENVIRONMENT Save the contents of the module module name into a file named filename If a file named filename already exists then the action take
27. AO 6 l CAPT PuntiOns esee REESE erre RE E deeded 40 6 2 Language Specific Interface 00 eee eee 58 6 2 1 TispeA PT etic wots ce R b 3 nbRIm eS RMEDUE ES UR Roe es 59 6 2 1 1 Common Lisp Initialization 2000 59 6 2 1 2 Type Declarations 0 0 59 0 2 1 3 NULL value8 eerte EROR ER clack ie ERRORS 59 6 2 1 4 Wrapped Literal Values 0 00 c cece eee 59 6 2 1 5 Special Variables 0 c cece eee 60 6 2 1 6 CLOS Objects versus Structs 0 00 cece ee ee eee 60 6 2 2 CHAPI cnt ces cate e eerte RU RU ERR DUC eee eas 60 6 2 3 Java APIs ke eade pK RARE ERR ERERPASRPEPIPEG S 60 6 2 8 1 Initializing PowerLoom 000 eee ee eee 60 6 2 3 2 PowerLoom Java Conventions 0 eee 60 6 2 8 3 Using the PLI Class 0 e eee 62 6 2 3 4 Using Stella Objects 0 cece eee eee 63 6 2 3 5 PowerLoom and Threads 000s 63 6 2 3 6 Setting and Restoring Global Variable Values 63 6 2 8 7 Java Character Mapping 00 cee eee eee ee 65 6 2 3 8 Stella Exceptions in Java 0c cece eee eee 66 6 2 3 9 Iteration in Java kk eee ees 66 6 2 3 10 Utility Classes for Java 0 eee ees 67 7 Built In Relations 68 8 PowerLoom GUI 84 8 amp 1 Invoking the GUL eed Eee E eei neces wl ete 84 8 1 1 Starting a PowerLoom Server 0 cece eee eee eee
28. BUSINESS retrieve all company name x y gt There are 3 solutions 1 X ACME CLEANERS Y ACME Cleaners LTD 2 X MEGASOFT Y MegaSoft Inc 3 X MEGASOFT Y MegaSoft Now observe the same query executed from within the alternate Business module in module ALTERNATE BUSINESS retrieve all company name x y Chapter 3 Annotated Example 23 gt There are 4 solutions 1 X ACME CLEANERS Y ACME Cleaners LTD 2 X MEGASOFT Y MegaZorch Inc 3 X WEB PHANTOMS Y Web Phantoms Inc 4 X MEGASOFT Y MegaSoft We see that all facts pertaining to company names have inherited down from the Business to the Alternate Business module except for the name for MegaSoft that we explicitly retracted Also the new facts asserted within the Alternate Business module appear mixed in with the inherited facts 3 13 Equality Reasoning PowerLoom makes the unique names assumption so every two different named logic con stants are assumed to be different For example assert Fred Joe gt Derived both TRUE and FALSE for the proposition P FALSE Clash occurred in module MDL PL KERNEL KB PL USER assert Fred Fred P TRUE However one can assert equality between skolems that represent function terms as well as between a function term skolem and a regular constant For example deffunction age x y assert age Fred age Joe assert age Fred 10
29. DALLAS 2 X TEXAS 3 X AUSTIN why 1 CONTAINS SOUTHERN US AUSTIN follows by Modus Ponens with substitution 11 SOUTHERN US 13 AUSTIN 12 TEXAS since 1 1 FORALL 11 713 lt CONTAINS 11 713 EXISTS 712 AND CONTAINS 11 712 CONTAINS 712 713 CONTAINS SOUTHERN US TEXAS CONTAINS TEXAS AUSTIN and 1 2 and 3 23 The following query combines a variety of relations that have been entered into the business modules It retrieves names of companies whose headquarters are in the southern US Note that query variables that do not appear in the output i e variables not listed after the all retrieve name exists city company and contains southern us city headquartered in company city company name company name There is 1 solution so far 1 NAME MegaSoft Inc why gt 1 and COMPANY NAME MEGASOFT MegaSoft Inc HEADQUARTERED IN MEGASOFT DALLAS CONTAINS SOUTHERN US DALLAS follows by And Introduction since 1 1 COMPANY NAME MEGASOFT MegaSoft Inc and 1 2 HEADQUARTERED IN MEGASOFT DALLAS and 1 3 CONTAINS SOUTHERN US DALLAS 1 3 CONTAINS SOUTHERN US DALLAS Chapter 3 Annotated Example 22 follows by Modus Ponens and substitution 13 DALLAS 12 TEXAS 11 SOUTHERN US since 1 3 1 forall 11 13 lt CONTAINS 11 13 exists 12 and CONTAINS 11 712 CONTAINS 12 13 CONTAINS SOUTHERN US TEXAS and 1 3 3 CONTAINS
30. EURO ER sees quate tate anne ee 12 3 4 Relation Hierarchies erscceeicicmiericed cinesi tinhon e 13 g0 BU icllOLs eeeeeee dedere e pxaer e head E 13 32 0 Defined Concepts iiic aee tu ret ARR Re ed ten dus 14 3 7 Negation and Open and Closed World Semantics 15 TE cere rn aa e eR OC 14 ded es edi edet E Pd 17 5 9 Clipping of Values cessere mh de pr PR seas 18 3 10 Rule based Inference 0 een eee ees 19 3 11 Explanation ides hens eR ga dhesen do beens eR aeange 20 3 12 Contexts and Modules 0 0 nirera esee 22 3 13 Equality Reasoning rne pdt ae RR MEA UAE 29 3 14 Classification Subsumption sese 23 3 15 ruth Maintenance cnc dictates bpEUNESEIHE Edere e es eng 24 3 16 Inference Control sssssseeeeee nn 24 3 17 Keyword Axioms 000 cece cece ni iie o oir 24 3 18 Cardinality Type Reasoning with Frame Predicates 24 3 19 Loom to PowerLoOm c i e RR E RE hey 24 3 20 Deviations from KIF sssssuleeeeee e 24 3 21 Differences from Loom 0 0 cece eren rren eens 24 3 22 Defalis sopies eii Vite eee ods bl ies DER eaten 24 3 23 Sets Lists SETOFALL KAPPA eeeeess 24 4 Communicating with PowerLoom 25 4 1 Command Interpreter 00 Ei SEE a eee ees 25 4 2 Persistent Knowledge Bases 0 c cece eee eens 26 5 CODIDIDADEGSg 244 t C CER E Poe eR 27 6 PowerLoom API sse
31. Fred PowerLoom can use a negation as failure proof strategy which returns TRUE if Fred cannot be proved to be happy Also if the re lation children is closed then a value for the expression range max cardinality children Fred can be inferred merely by counting the number of fillers of the children role on Fred COLLECT INTO ASCENDING SET c COLLECTION sortby RELATION Function gt orderedset LIST Collect elements of c into an ascending orderedSet where the position of each ele ment is determined by the value computed for it by the sortBy relation Ordering is done via sorting as opposed to using a comparison relation similar to the sort by option to the retrieve command If sortBy is not single valued the position of an element is determined by its largest sortBy value Note that similar to other collect into functions c can be a named concept a setofall or an enumer ated collection and sortBy can be a named relation or a kappa COLLECT INTO DESCENDING SET c COLLECTION sortby RELATION Function gt orderedset LIST Collect elements of c into a descending orderedSet where the position of each ele ment is determined by the value computed for it by the sortBy relation Ordering is done via sorting as opposed to using a comparison relation similar to the sort by option to the retrieve command If sortBy is not single valued the position of an element is determined by its largest sortBy
32. NOT cycle through all relations searching for ones that permit the proof to succeed However users can obtain the same effect if they choose by using other second order predicates to generate relation bindings For example the query retrieve all x and Relation r holds r Fred x retrieves all constants for which there is some binary relation that relates Fred to that relation Chapter 7 Built In Relations 74 IMAGE URL x THING url STRING Relation url is a URL pointing to an image illustrating x The Ontosaurus browser looks for image url values attached to objects it is presenting and displays them prominently thereby spiffing up its displays INEQUALITY x THING y THING Relation Abstract superrelation of inequality relations INSERT ELEMENT l LIST n INTEGER e THING gt r LIST Function Add e at position n zero based to 1 to construct r shifts the remaining elements right Count from end of the list and shift left if n is negative such that 1 inserts at the end of the list 2 second to last etc INSTANCE OF x THING c COLLECTION Relation True if x is an instance of c Can be used to generate concept values of c given an instance x INVERSE r BINARY RELATION gt inverserelation THING Function Function that returns the inverse relation for r PERFORMANCE NOTE for best results there should be only one inverse R I assertion per relation pair R a
33. Note the use of bound variables to screen out unbound variables before they are passed to the setofall term COLLECTION ABSTRACT COLLECTION Class Not documented any value OBJECT Class Parameter of Not documented COLLECTIONOF m THING gt c COLLECTION Function Abstract function existing to subsume SETOF and LISTOF COMMENT x THING 7s STRING Relation 7s is a comment attached to x Comments are a generalization of other annotations such as documentation and issue strings COMMUTATIVE r RELATION Relation A relation r is commutative if its truth value is invariant with any permutation of its arguments CONCEPT x RELATION Concept The class of reified unary relations The Powerloom notion of concept corresponds to the object oriented notion of class From a logic standpoint the notion of a concept is hard to distinguish from the notion of unary relation The conceptual distinction is best illustrated in the domain of linguistics where concepts are identified with collective nouns while unary relations are identified with adjectives For example Rock is a concept while rocky is a unary relation CONCEPT PROTOTYPE c CONCEPT gt i THING Function Function that given a concept returns a prototypical instance that inherits all con straints that apply to any concept member and has no additional constraints Chapter 7 Built In Relations 71 CONTEXT THING Class Not documented
34. Return true if the STELLA feature is currently disabled disjoin truth values tvl TRUTH VALUE tv2 TRUTH VALUE Function TRUTH VALUE Return the logical disjunction of truth values tv1 and tv2 disjoint terms dl DESCRIPTION d2 DESCRIPTION BOOLEAN Function Return TRUE if d1 and d2 belong to disjoint partitions do clear instances module MODULE Function Function version of clear instances that evaluates its argument do save module module MODULE store OBJECT Function Save module to the persistent store store which can either be an output stream or a persistent OBJECT STORE empty self BOOLEAN Method on QUERY SOLUTION TABLE Return TRUE if self has zero entries empty self BOOLEAN Method on FLOAT VECTOR Return TRUE if self has length 0 enabled powerloom feature feature KEYWORD BOOLEAN Function Return true if the STELLA feature is currently enabled estimated length self INTEGER Method on PAGING INDEX Return the estimated length of the sequences in self which could be too large if some of the members have been deleted Chapter 11 Miscellaneous 118 evaluate proposition self PROPOSITION Function Evaluate self against its arguments possibly resulting in the setting or changing of its truth value evaluation state proposition PROPOSITION KEYWORD Function Return POSTED if proposition is on the evaluation queue for context EVALU ATED if has been e
35. Rj Concepts for AIRCRAFT KB i Relations for AIRCRAFT KB f Instances for AIRCRAF e n ROOT MODULE J CONCEPT ROOT AD RELATION ROOT veer ae ave a H POWERLOOM SERVER C ACTIVE GUIDANCE TYPE ATTACKED BY TF 111 9 H PL KERNEL KB ENTITY CALIBER I F 117 H CYC FRAME ONTOL c PHYSICAL ENTITY CARRIED BY I F 118 GE 100 M KIF FRAME ONTOLO c SOCIAL ENTITY CEILING IFS H LOOM API GROUND UNIT COST 9 H PL USER A GUIDANCE TYPE CREW ROLES IF 16 H AIRCRAFT KB P MISSION CREW SIZE IF100 PW 100 A PASSIVE GUIDANCE TYPE DESIGNATION IF110 GE 200 TECHNOLOGY DIAMETER E C ALL WEATHER ENGINE COUNT E EHELCOW CIO C STEALTH TECHNOLOGY ENGINE TYPE j 1 F404 C SWING WING ENGINE USED BY I FAIRCHILD REPUBLIC WIDTH RELATION FLEET SIZE I GARRETT FUEL CAPACITY d rcr GUIDANCE HAS MISSION i I GENERAL DYNAMICS nhcm7mmocxmmmmmmmzoemn Propositions for F 15 View From Module AIRCRAFT KB Y EY P ALL WEATHER F 15 a P CARRIED BY 20MM VULCAN F 15 P CARRIED BY 600G TANK F 15 P CARRIED BY AIM 120 F 15 P CARRIED BY AIM 7M F 15 hd View From Module AIRCRAFT KB v Figure 8 3 Knowledge Browser The browser consists of several subpanes which we refer to as navigation panes Each navigation pane consists of a title pane a toolbar and a content pane The title pane contains a title indicating what is displayed in the content pane The toolbar consists of zero or more buttons which per
36. a list of all definite TRUE or FALSE propositions that reference the instance instanceRef Chapter 11 Miscellaneous 113 call ask query OBJECT TRUTH VALUE Function Callable version of ask which see Accepts queries specified by a query iterator or specified as a CONS list of arguments as they would be supplied to ask Raises LOGIC EXCEPTIONSs in case of illegal queries and logical expressions call defconcept arguments CONS NAMED DESCRIPTION Function Callable version of the defconcept command which see Expects the same argu ments as defconcept but supplied as a list call deffunction arguments CONS NAMED DESCRIPTION Function Callable version of the deffunction command which see Expects the same argu ments as deffunction but supplied as a list call defobject arguments CONS LOGIC OBJECT Function Callable version of the defobject command which see Expects the same arguments as defobject but supplied as a list call defproposition arguments CONS PROPOSITION Function Callable version of the defproposition command which see Expects the same arguments as defproposition but supplied as a list call defrelation arguments CONS NAMED DESCRIPTION Function Callable version of the defrelation command which see Expects the same argu ments as defrelation but supplied as a list call list undefined relations module MODULE Function local BOOLEAN CONS Callable version of l
37. and KEYWORD Unless you are in a case sensitive module the following four commands are equivalent all facts of Merryweather all facts of MERRYWEATHER all facts of merryweather all facts of MerryWeather Commands can also have amp rest parameters similar to Lisp functions These are either used to allow a variable number of arguments or to handle optional arguments since STELLA does not directly support optional arguments Here is a list of important parameter types used in the command specifications below e GENERALIZED SYMBOL A generalized symbol is either a plain symbol similar to a Lisp symbol such as Merryweather a keyword similar to a Lisp keyword such as KIF or a STELLA surrogate which is a symbol starting with an at sign e g CONS STELLA surrogates are used as names for objects of arbitrary types e NAME Names can be either a string or a GENERALIZED SYMBOL i e a symbol a keyword or a surrogate If a symbol is supplied only its symbol name is used Commands that take names as arguments usually coerce whatever argument is entered into a string but by allowing a NAME they make it a little bit more convenient to type a name in an interactive invocation e PARSE TREE A parse tree is similar to a Lisp s expression i e it can either be an atom such as a symbol number or a string or a list of zero or more parse trees For example the expression happy Fred is a parse tree and so are its compon
38. and those modules below it is the case that Georgia is a state If we want the fact that Georgia is a state to be recognized as true in many or most other modules then we should make our assertion in a module that is relatively high up in the hierarchy so that is visible to inherited by the other modules The inheritance of facts is not monotonic a child module can retract or override facts inherited from its ancestors For example suppose we have two modules called above and below such that the below module is below inherits from the above module Next suppose we make an assertion within the above module that Joel is a duck and then we shift to the below module and retract the proposition that Joel is a duck From the vantage point of the below module if we now ask if Joel is a duck we will get back the value unknown However if we switch to the above module and ask the same question we get back the answer true This occurs because the effect of the retraction operation that was applied to the below module is not visible to modules above it or to any sibling modules Hence when module hierarchies are involved it is oversimplifying to state that a retraction has the effect of erasing a prior assertion The PowerLoom execution process maintains a pointer to the current module and all asserions queries etc are made relative to that module Hence when we talk about Chapter 2 Conceptual Framework 10 switchi
39. arguments that have been asserted or inferred by forward chaining relation and arguments is a string that begins with a left parenthesis followed by a relation name one or more argument identifiers and terminated by a right parenthesis Each argument identifier can be the name of a logical constant a literal reference e g a number the null identifier or a variable an identifier that begins with a question mark Each occurrence of a null or a variable acts like a wild card A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get propositions of object name STRING module name STRING Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return all propositions that have the object named object name among their argu ments and that are TRUE in the scope of the module module name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get range relation name STRING module name STRING Function environment ENVIRONMENT LOGIC OBJECT Return the type a concept for fillers of the binary relation relation name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such contex
40. be changed by using the set method Note that we use variables of type Object to hold the values since that avoids the need to cast when extracting the current value since the only operation we do with the current value is save it to restore it later import edu isi stella CONTEXT CHANGE Object savedModule Stella MODULE get Module newModule Stella getStellaModule contextName true if newModule null Handle missing module J try Module changeCurrentModule newModule Code that uses the newModule finally Module changeCurrentModule savedModule INTEGER VALUE CHANGE Object savedValue Stella SAFETY get try Stella SAFETY set new Integer 3 Code that uses the newModule finally Stella SAFETY set savedValue BOOLEAN VALUE CHANGE Object savedValue Stella PRINTREADABLY get try Stella PRINTREADABLY set Boolean TRUE Chapter 6 PowerLoom API 65 Code that uses the newModule finally Stella PRINTREADABLY set savedValue The need to change the module using this type of code can be largely avoided by using the functions in the PLI interface package They take a module argument and can handle the binding and restoration of the module value themselves 6 2 3 7 Java Character Mapping The full Stella to Java character mapping is the following The character mappings use uppercase characters if the basic identifier
41. bound RANGE MAX CARDINALITY r RELATION i THING gt Function maxcard INTEGER Returns the strictest computable upper bound on the cardinality of the set of fillers of the relation r applied to i see the discussion for range cardinality lower bound RANGE MIN CARDINALITY r RELATION i THING gt Function mincard INTEGER Returns the strictest computable lower bound on the cardinality of the set of fillers of the relation r applied to i see the discussion for range cardinality lower bound RANGE TYPE r RELATION i THING type COLLECTION Relation Relation that specifies a type range of the relation r applied to i Multiple range types may be asserted for a single pair r i Technically a retrieval of types for a given pair should include all supertypes superconcepts of any type that is produced but for utility s sake only asserted or directly inferrable types are returned REFLEXIVE r RELATION Relation A binary relation r is reflexive if it is always true when both of its arguments are identical REFUTATION QUERY prop PROPOSITION Relation Query prop with inference level set to refutation Equivalent to query prop inference level refutation but more efficient RELATION MAPPABLE OBJECT Class Not documented abstract BOOLEAN Slot of Not documented RELATION COMPUTATION r RELATION Relation computation COMPUTED PROCEDURE Names a compu
42. concept needs to be created separately Chapter 6 PowerLoom API 53 A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly s create function name STRING arity INTEGER Function module name STRING environment ENVIRONMENT LOGIC OBJECT Create a function named name with arity arity in the designated module Domain and range information can be added via assertions of domain nth domain and range relations A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly s create object name STRING concept name STRING Function module name STRING environment ENVIRONMENT LOGIC OBJECT Create an object named name of type concept name in the designated module Both name and concept name can be null strings If name is a null string then an object will be created with a new non conflicting name based on concept name or system generated if no concept nameis specified If concept name is the null string then the o
43. context exception is thrown s clear module name STRING environment ENVIRONMENT Function MODULE Destroy the contents of the module named name as well as the contents of all of its children recursively If no module named name exists a Stella no such context exception is thrown s conceive sentence STRING module name STRING Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Create one or more proposition objects from the sentence sentence in the module named module name Return an iterator of the propositions If any of the new propo sitions has the same structure as an already existing proposition an automatic check for duplicates will return the pre existing proposition Multiple propositions may be returned for a single sentence because of normalization of equivalences conjunctions etc A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella No Such Context Exception is thrown Signals a Proposition Error if PowerLoom could not conceive sentence s create concept name STRING parent name STRING Function module name STRING environment ENVIRONMENT LOGIC OBJECT Create a concept named name in the designated module with with the concept named parent name as superconcept which can be left undefined Additional superconcepts can be added via assertions of the subset of relation Note that a specified parent
44. cpp function code a pointer to the function code Common Lisp FUNCTION result of ZZ or FUNCTION Java java lang reflect Method register specialist function name STRING Function function reference FUNCTION CODE module MODULE environment ENVIRONMENT Register name as a function name in module which will invoke the native code pro cedure described by function reference The name is a fully qualified name which will be interpreted by the normal rules for reading names in PowerLoom The function must conform to the signature for specialist functions The exact form of function reference depends on the underlying programming lan guage The following type mappings are used C Common Lisp FUNCTION result of or FUNCTION Java java lang reflect Method s register computation function name STRING Function native name STRING arity INTEGER module name STRING environment ENVIRONMENT Register name as a function name in the module named module name This function will the native code named native name The name is a fully qualified name which will be interpreted by the normal rules for reading names in PowerLoom The native name will be processed in a manner that depends on the underlying programming language The following type mappings are used C Not available Error signaled Common Lisp The native name is read by READ FROM STRING and then the SYMBOL FUNCTION is taken Java A fully package quali
45. different port Once the Ontosaurus server is running go to your Web browser and enter the following URL http localhost 9090 ploom ontosaurus Substitute a fully qualified host name and different port if necessary Ontosaurus comes with a short online help page and is otherwise self explanatory Alternatively the PowerLoom Web Archive available in powerloom x y z native java lib ploom war can be deployed in an Apache Tomcat server by copying it into the server installation s webapps directory Ontosaurus can then connect to it at the above URL at the server port used by Tomcat Chapter 10 Installation 104 10 Installation 10 1 System Requirements To install and use PowerLoom you ll approximately need the following amounts of disk space e 16 MB for the tarred or zipped archive file e 70 MB for the untarred sources tanslations compiled Java files and documentation e 16 MB to compile a Lisp version e 16 MB to compile the C version without g e 5 MB to compile the Java version already included This means that you will need approximately 100 MB to work with one Lisp one C and one Java version of PowerLoom in parallel If you also want to experiment with the Lisp translation variant that uses structures instead of CLOS instances to implement STELLA objects then you will need an extra 16 MB to compile that The full PowerLoom development tree is quite large since for every STELLA source file there are three to fou
46. eiaa i a 105 10 3 Lisp Installation isse bee rr et eR RR xd 105 10 4 C Installation 0 cect eee eens 106 10 5 Java Installation 0 ccc eee cee e eee e eens 107 10 6 Removing Unneeded Files 0 e cece eee eens 107 10 7 Installing PowerLoom Patches 0000 e eee ee eee 108 11 Miscellaneous 0 109 12 Glossary T 134 13 PowerLoom Grammar 137 13 1 Alphabet hber eret ate ei nde tee wee eels 137 13 2 Grama edit Mete ate tates Gains cate Ll aN 137 13 2 1 Constants and Typed Variables 0000 137 193 2 2 EGFInS sas caddis ertt ert weeny Pan cdd Ree a 138 13 2 3 Sentenees s cescess reve e ODIO RORESR EE S edens 138 13 244 Defmitiongs iiic 4 repe em CR LESS DEDE ERENREPEMES 139 Function Index oru RR bie e pedis 141 Variable IileX cu 52 nex dno iex ier SEO Rx ene Ead 142 Concept Index coi ooo ret ee een Ner CH 143 ii Chapter 1 Introduction 1 1 Introduction This document describes the PowerLoom knowledge representation and reasoning system PowerLoom is the successor to the Loom knowledge representation system It provides a language and environment for constructing intelligent knowledge based applications Pow erLoom uses a fully expressive logic based representation language a variant of KIF It uses a natural deduction inference engine that combines forward and backward chaining to derive what logically follows from
47. entirely semantic However some operators will only accept specific constants In contrast to the specification of KIF3 0 PowerLoom supports a typed syntax There fore variables in quantified terms and sentences can appear either typed or untyped as follows lt vardecl gt lt indvar gt lt constant gt lt indvar gt 13 2 2 Terms Terms are used to denote objects in the world being described term lt indvar gt constant lt funterm gt lt listterm gt lt setterm gt lt quanterm gt lt listterm gt listof lt term gt lt setterm gt setof lt term gt lt funterm gt lt funconst gt lt term gt Note Zero arguments are allowed for lt funterm gt in KIF3 0 lt term gt lt quanterm gt the lt vardecl gt lt sentence gt setofall lt vardecl gt lt sentence gt kappa lt vardecl gt lt vardecl gt lt sentence gt lambda lt vardecl gt lt vardecl gt lt term gt Note KIF3 0 allows term instead of lt vardecl gt for setofall No lt quanterm gt as well as no lt setterm gt in core of KIF as a result of descision 95 3 March 1995 13 2 3 Sentences Sentences are used to express propositions about the world lt sentence gt constant equation inequality lt relsent gt lt logsent gt lt quantsent gt lt equation gt lt term gt lt term gt lt inequality gt lt term gt lt te
48. even unknown in case the propositon was just newly constructed RELATION SPECIALIST r RELATION sp COMPUTED PROCEDURE Relation Names a specialist a function that evaluates an atomic relation proposition dur ing query processing This defines an extensible means for computing with the control stack The function is passed a CONTROL FRAME that contains the proposition and returns a keyword FINAL SUCCESS CONTINUING SUCCESS FAILURE or TERMINAL FAILURE that controls the result of the computation SCALAR x SCALAR Concept The class of scalar quantities SCALAR INTERVAL x SCALAR Concept An interval of scalar quantities SECOND ELEMENT COLLECTION gt e THING Function Return the second element e of SET LIST SET MIXIN Class Not documented any value OBJECT Class Parameter of Not documented SETOF m THING gt c SET Function Term forming function that defines an enumerated set consisting of all function ar guments setof is like Listof except that it removes duplicate values SHALLOW QUERY prop PROPOSITION Relation Query prop with nference level set to shallow Equivalent to query prop inference level shallow but more efficient SINGLE VALUED c RELATION Relation The relation c is single valued if the value of its last argument is a function of all other arguments All functions are single valued see function SQUARE ROOT x NUMBER y NUMB
49. have Java installed on your system these scripts should run out of the box without any further installation requirements If you want to use the Lisp version of PowerLoom simply load the file 1oad powerloom lisp into your Common Lisp If you want to use the C version you have to compile it first See the Installation section in this manual for more details on how to install the Lisp C or Java version of PowerLoom See Chapter 10 Installation page 104 Under Unix or MacOS X open a shell window somewhere to run PowerLoom For example powerloom Running Java version of PowerLoom Initializing STELLA Initializing PowerLoom Welcome to PowerLoom 4 0 0 Copyright C USC Information Sciences Institute 1997 2010 PowerLoom comes with ABSOLUTELY NO WARRANTY Type copyright for detailed copyright information Type help for a list of available commands Type demo for a list of example applications Type bye exit halt quit or stop to exit PL USER Under Windows you can do something similar by running a Command Prompt window and executing the powerloom bat script You can also simply double click on the script which will run PowerLoom and bring up a Command Prompt window for you Chapter 1 Introduction 5 Once the prompt has come up you can type in PowerLoom commands and see their results The string preceding the prompt indicates the current module
50. in STELLA Finally if you don t need any of the STELLA sources you can delete the directory sources stella If you don t need local copies of the STELLA and PowerLoom documentation you can delete parts or all of the directories sources stella doc and sources logic doc 10 7 Installing PowerLoom Patches If you already have an older version of PowerLoom installed you can upgrade to the latest patch level by downloading incremental sets of patches instead of downloading the complete release Patch files are usually significantly smaller than the complete release and patch ing an existing installation can also preserve local additions deletions or modifications PowerLoom patch files are available from the same location as the full PowerLoom release Patches are currently only available in Unix diff format which requires the Unix patch utility to install them the patch program should be readily available on most Unix systems otherwise you can get it for free from the Free Software Foundation Patch files follow the following naming convention Suppose the current version of PowerLoom is 3 0 0 Then the patch file to update to the next patch level is called powerloom 3 0 0 3 0 1 diff gz Important Patch files are strictly incremental Thus to upgrade from version 3 0 0 to 3 0 2 you will need two patch files one to go to version 3 0 1 and one to go from that to version 3 0 2 you will have to apply them in that
51. instances will not be copied to the target module therefore the resulting propositions might reference external out of module objects in case they are not visible there If record justifications is TRUE justifi cations will be recorded for the query and the resulting justifications will be linked to the asserted propositions Here are some examples assert from query retrieve all foo x y assert from query retrieve all y x exists z and foo x z foo z y relation bar module other assert from query retrieve all and relation x symmetric x pattern kappa pred forall x y gt holds pred x y holds pred y x assert rule ruleName NAME PROPOSITION N Command Set the truth value of the rule named ruleName to TRUE The proposition having the name ruleName may be any arbitrary proposition although we expect that it is probably a material implication See retract rule cc amp rest name NAME CONTEXT N Command Change the current context to the one named name Return the value of the new current context If no name is supplied return the pre existing value of the current context cc is a no op if the context reference cannot be successfully evaluated classify relations module NAME local BOOLEAN N Command Classify named relations visible in module If local only classify descriptions defined within module i e don t classify descriptions inherited from
52. is an enumerated collection This subsumes both sets and lists is enumerated list obj OBJECT BOOLEAN Function Test whether obj is an enumerated list is enumerated set obj OBJECT BOOLEAN Function Test whether obj is an enumerated set is false tv TRUTH VALUE BOOLEAN Function Tests whether tv is a false truth value It can be false either absolutely or by default is float obj OBJECT BOOLEAN Function Test whether obj is of type FLOAT double is inconsistent tv TRUTH VALUE BOOLEAN Function Tests whether tv is an inconsistent truth value is integer obj OBJECT BOOLEAN Function Test whether obj is of type INTEGER is logic object obj OBJECT BOOLEAN Function Test whether obj is of type LOGIC OBJECT is number obj OBJECT BOOLEAN Function Test whether obj is of type NUMBER This can be either an integer or a floating point number One key characteristic is that object to integer and object to float will both work on it is strict tv TRUTH VALUE BOOLEAN Function Tests whether tv is a strict non default truth value is string obj OBJECT BOOLEAN Function Test whether obj is of type STRING Chapter 6 PowerLoom API 49 is subrelation sub LOGIC OBJECT super LOGIC OBJECT Function module MODULE environment ENVIRONMENT BOOLEAN Return TRUE if sub is a subconcept subrelation of super is true tv TRUTH VALUE BOOLEAN Function Tests whether tv i
53. must have the correct type In particular integer values will not be coerced to floating point values by the code The following native type assignments are made Stella Common Lisp INTEGER FIXNUM FLOAT DOUBLE FLOAT STRING SIMPLE STRING For convenience loading PowerLoom will set the default format for reading floating point numbers in Common Lisp to be double float Stella CONS objects are implmented as native Lisp conses Boolean values can take on the values stella true or stella false 6 2 1 3 NULL values One additional consequence of the strong typing of the language is that there are specialized NULL values for numeric and string parameters Stella Type Null Value INTEGER stella null integer FLOAT stella null float STRING stella null string 6 2 1 4 Wrapped Literal Values Literal values integers floats strings etc that are used in PowerLoom appear as wrapped values The PowerLoom API functions object to can be used to coerce the values into the appropriate return type Chapter 6 PowerLoom API 60 lt to be written wrapping values gt 6 2 1 5 Special Variables All Stella special variables are implemented as Common Lisp special variables Binding of the values can be used normally 6 2 1 6 CLOS Objects versus Structs PowerLoom can be translated in one of two ways for Common Lisp One method uses CLOS objects as the basis for all Stella and PowerLoom objects For faster execution it is also
54. nt iseesniescosdbh y dade 43 get concept cgcka e s hee pre ee te be 44 get concept instance matching value 44 get concept instances lesse 44 get concept instances matching value 44 get current module 000 ini 44 get direct concept instances 44 get direct subrelations sess 44 get direct superrelations 44 get dir ct types onooss n eer 44 get domaln eee ni ninian EER bert 44 get enumerated collection members 44 get forward justifications 119 get hom modul6 e isle bore P RARE 45 get inferred binary proposition values 45 get instanc i s e des cade eddie ra eed 119 get keyWOId ise uusads adea er e be 131 get load path i iididean ed die so Rr khoa 34 get module s sss ineo d estpeubdur 45 119 get modulesS 2 duco ie ERE REA RARE bad sce bee 45 143 get name in module reae n eng 131 get ntb d omain 2 kde di e ER ERE 45 get nth float geapouee ek en Enea 45 get ntb integer cessa da mener 45 get nth logic object cies csse eee eas 46 get nth String i 6 Ryu dogARREPS SERERE 46 get nth valu 2i legc n heb RIPE 6 46 get objJeCtiiio d vea uu eiua KG ed dde 46 get Operator i ix leis eau pn RR ead 46 get parent modules 0 eae 46 get predicate o ebbe e REPEPE rsk 46 get proper su brelations s AT get proper superrelations
55. object to parsable string self OBJECT STRING Function Return a string representing a printed representation of the object self Like object to string but puts escaped double quotes around strings object to string self OBJECT STRING Function Return a printed representation of the term self as a string print rules relation OBJECT N Command Print the list of true rules associated with relation reset powerloom Function Reset PowerLoom to its initial state CAUTION This will destroy all loaded knowl edge bases and might break other loaded STELLA systems if they do reference Pow erLoom symbols in their code retract proposition PARSE TREE OBJECT N Command Retract the truth of proposition Return the retracted proposition object KIF example retract happy Fred retracts that Fred is happy Note that for this assertion to succeed the relation happy must already be defined If the constant Fred has not yet been created it is automatically created as a side effect of calling retract retract binary proposition relation LOGIC OBJECT arg OBJECT Function value OBJECT module MODULE environment ENVIRONMENT PROPOSITION Retract that the proposition relation arg value is TRUE in module Return the asserted proposition retract nary proposition relation and arguments OBJECT Function module MODULE environment ENVIRONMENT PROPOSITION Retract the proposition that arguments satis
56. on expressive representa tion we have not yet focused much on Semantic Web languages such as OWL which restricts expressivity to gain decidability OWL also has other limitations such as re Chapter 1 Introduction 4 striction to binary relations and lack of support for arithmetic and inequalities which limits its usefulness for practical applications Nevertheless given that people are starting to use these languages more and more we ve developed some initial import translators for RDF RDFS and OWL which once they mature we will release as part of PowerLoom e Portability and integration Since PowerLoom is implemented in STELLA it is avail able in Lisp C and Java implementations and highly portable PowerLoom can be run in each of these languages under Unix such as Linux SunOS or MacOS X as well as Windows 2000 and XP Due to the availability in three main stream languages it can easily be integrated programmatically with most application software without having to use some complex integration framework The C and Java libraries for PowerLoom are also quite small and light weight for example the necessary STELLA and PowerLoom jar files of the Java implementation are less than 2 5 Meg in size 1 2 Powerloom History lt to be written gt 1 3 Running PowerLoom The easiest way to run PowerLoom on a variety of platforms is to use the powerloom or powerloom bat scripts in the top level PowerLoom directory If you
57. relations instances and propositions rules which are described below Module Editor The module editor contains a number of fields and components used to enter information relevant for a new or existing module Examples of values that can be edited are a module s name its documentation and its includes or supermodules list Concept Editor The concept editor allows editing of concept attributes such as a con cept s supertypes its name its associated propositions etc In addition to the inherent attributes of a concept all relations which have the concept as a domain type are displayed and may be edited Clicking the button above the relation list opens a new relation editor with default values filled in Similarly clicking the button above the proposition list opens a proposition editor Chapter 8 PowerLoom GUI 95 Relation Editor The relation editor allows the user to enter a list of variables and types for the relation s arguments and allows the setting of various relation attributes e g whether the relation is closed functional etc Similar to the concept editor propositions and rules associated with a relation can be edited Instance Editor The instance editor allows the user to input an instance s name doc umentation and associated propositions If a proposition uses the relation image url an image will be retrieved from the server and presented in the editor window Proposition editor The proposition edit
58. retrieve all Person x retrieve all x Person x will find all people Note that in the first case we only specify a query variable and its type but omit the logic sentence which defaults to TRUE This somewhat impoverished looking query can be paraphrased as retrieve all x of type Person such that TRUE Chapter 5 Commands 38 retrieve x or happy x parent of Fred x will try to find a person that is happy or has Fred as a parent retrieve y x parent of x 7 y will try to find the one pair of parent child and return it in the order of child parent retrieve all x Person exists y Person parent of x y will generate the set of all parents Note that for these queries to run the class Person the relations happy and parent of and the logic constant Fred must already be defined see assert Use set unset feature trace subgoals to en disable goal tracing of the infer ence engine save module name NAME file STRING Command Save all definitions and assertions of module name to file set feature amp rest features NAME LIST OF KEYWORD N Command Enable the PowerLoom environment feature s named by features Return the list of enabled features Calling set feature without any arguments can be used to display the currently enabled features The following features are supported just in time inference Enables interleaving of forward chaining inference within backward chaining qu
59. sequence To find out the current version of your PowerLoom installation look at the version string displayed when the powerloom function is called in Lisp or when the C or Java program starts up Important Before you apply any patches you should always make a backup copy of your current PowerLoom installation to preserve any local modifications you made in case something goes wrong To apply patches copy the appropriate patch file to the top level of your PowerLoom installation directory Suppose the patch file is called powerloom 3 0 0 3 0 1 diff gz You can apply the patches using the following command gunzip qc powerloom 3 0 0 3 0 1 diff gz patch p1 If you deleted some native PowerLoom files to save space you can use the f option to force patch to proceed even if files it needs to patch do not exist anymore this is somewhat dangerous in case patch encounters some more serious problems For example gunzip qc powerloom 3 0 0 3 0 1 diff gz patch p1 f To keep patch files small PowerLoom patch files do not contain updated binary files that ship with the full release such as Java class files jar files and PDF documents Those have to be regenerated either manually or with help of available Makefile s The Lisp version of PowerLoom will automatically recompile the first time the updated system is loaded To recompile the C installation follow the installation instructions given above The Java version wil
60. since they are implemented on a per thread basis The most important special variable is the reasoning context See Section 6 2 3 6 Setting and Restoring Global Variable Values page 63 6 2 3 6 Setting and Restoring Global Variable Values As noted above special variables in Stella are implemented as static fields in a catchall class named the same as the Stella module It will be in the java package corresponding to that Stella module The values of Special variables are stored in Java objects of the type StellaSpecialVariable a subclass of Java s InheritableThreadLocal Any changes Chapter 6 PowerLoom API 64 made to the values will not affect any other running threads This means that the changes don t need to be synchronized Note that global as opposed to special variables don t use these objects Numbers and boolean values are stored in special variables using the corresponding Java classes Integer Double Boolean etc The naming convention is to have all upper case letters with a dollar sign at the beginning and end of the name To temporarily change the value of one of these variables users will need to be responsible for saving and restoring the old values Use of the try finally construct is very useful for this since it guarantees that the restore of values will be done An example follows of how to safely change modules Contexts should be changed using the functions although other global variables can
61. the case of a binary relation means that a single first value can be mapped by the relation to more than one second value In the present case our model permits a company entity to have more than one company name If a binary relation always maps its first argument to exactly one value i e if it it single valued we can specify it as a function instead of a relation For example we can use a function to indicate the number of employees for a company deffunction number of employees c company gt n INTEGER When defining a function all arguments but the last appear just as they do for a relation The last argument and its type appears by itself following the keyword gt Defining a single valued relation as a function allows us to refer to it using a functional syntax within a logical sentence as in the following assert number of employees ACME cleaners 8 assert number of employees megasoft 10000 Chapter 3 Annotated Example 14 The functional syntax often results in shorter expressions than equivalents that use relational syntax For example to retrieve all companies with fewer than 50 employees we can simply write retrieve all and company x lt number of employees x 50 There is 1 solution 1 X ACME CLEANERS Using the syntax for relations the same query would require the introduction of an existential quantifier as in retrieve all and company x exists n and numb
62. the edit toolbar button can be pressed Object editors do double duty as object viewers since all relevant object information is present in each editor There are several common user actions that are available in edit dialogs For example hitting RET while the cursor is positioned in the name field of the editor commits the concept Most editors have OK and Cancel buttons at the bottom to commit or abort edits Lists of items commonly have and buttons at the top that allow one to add a new item or delete the selected item When the button is pressed either a chooser dialog see Section 8 4 6 Chapter 8 PowerLoom GUI 94 Choosers page 95 or a specialized editor will be opened Similar to the browser list items can be right clicked to display a list of possible actions For example a superconcept can be clicked in a concept editor to immediately edit the concept s parent 1 Editing Instance F 15 C ALL WEATHER C FIXED WING AIRCRAFT Documentation he F 15 Eagle is an all weather extremely maneuverable tactical fighter designed to gain and maintain air superiority in aerial combat Module AIRCRAFT KE Propositions HA P ALL WEATHER F 15 P CARRIED BY 20MM VULCAN F 15 P CARRIED BY 600G TANK F 15 e ean Figure 8 4 Instance Editor Each type of object has a specialized editor For example an instance editor is shown in Figure 8 4 There are separate editors for modules concepts
63. the facts and rules asserted in the knowledge base While PowerLoom is not a description logic it does have a description classifier which uses technology derived from the Loom classifier to classify descriptions expressed in full first order predicate calculus PowerLoom uses modules as a structuring device for knowledge bases and ultra lightweight worlds to support hypothetical reasoning To implement PowerLoom we developed a new programming language called STELLA which is a Strongly Typed Lisp like LAnguage that can be translated into Lisp Java and C STELLA tries to preserve those features of Lisp that facilitate symbolic programming and rapid prototyping while still allowing translation into readable as well as efficient Java and C code Because PowerLoom in written STELLA we are able to deliver it in all three languages 1 1 Powerloom Features PowerLoom is a full function logic based knowledge representation and reasoning system that supports all aspects of knowledge based applications It allows the representation of complex knowledge in a declarative logic based language supports a variety of reasoning mechanisms to make implicit knowledge explicit has a powerful query engine to retrieve what has been asserted and logically follows from the KB provides file based and RDBMS based storage of knowledge bases has a context and module system to effectively partition and organize large knowledge bases and has an extensive API in
64. time the application starts up 8 4 2 Edit Preferences A preferences dialog can be opened by selecting the Edit gt Edit Preferences menu item Currently the only preference that a user can edit is whether or not to open a browser when the application is started The dialog contains a checkbox asking whether or not the preferences should be saved If the checkbox is not checked the preferences will remain in effect for the duration of the current session but will not be in effect when the application is restarted 8 4 3 KB Load Save In its standard configuration PowerLoom stores knowledge bases via flat files The user can load one or more KB files from the local file system which will be transferred across the network and loaded into the server instance via PowerLoom s load command Conversely a modified knowledge base can be saved to a local file on a per module basis using a version of PowerLoom s save module command Future versions of PowerLoom will support more sophisticated persistence that will go directly to a database or other persistent store 8 4 4 Browsing The knowledge browser window shown in Figure 8 3 can be opened by selecting the KB Browse menu item or typing Ctrl b The browser provides a visual overview of all knowledge in the KB and is capable of launching specialized tools such as editors search dialogs etc Chapter 8 PowerLoom GUI 91 E Browsing Knowledge Base blackcatisi edu 8000 Modules
65. we assert that ZZ Productions is not an S corporation then PowerLoom knows that the proposition in question is false assert not s corporation zz productions ask s corporation zz productions FALSE ask not s corporation zz productions TRUE After asserting that ZZ Productions is not an S corporation a repeat of the query asking if it is one will now return false because the explicit assertion of the negation allows a quick disproof of the positive query Note PowerLoom uses all its effort to prove that the proposition in question is true and only uses some effort to prove that it is false Therefore only falsities that are discovered on the way or with shallow inference strategies will be found which was the case above If you want to check whether a proposition is false with maximum effort simply ask the 2 Actually PowerLoom implements a five valued logic the remaining two values are default true and default false However the present discussion defers the subject of default truth values 3 Because proving negations can be very difficult PowerLoom will only conduct a very quick and shallow search for a disproof More extensive reasoning is used if a negation is asked about explicitly thus PowerLoom may return unknown if asked about P but true if asked about not P Chapter 3 Annotated Example 16 negated proposition by wrapping an explicit not arount it The reason for this asymme
66. we can ask why Dallas is contained in the Southern US ask contains southern us dallas TRUE why gt 1 CONTAINS SOUTHERN US DALLAS follows by Modus Ponens and substitution 13 DALLAS 12 TEXAS 11 SOUTHERN US since 1 1 forall 11 713 lt CONTAINS 11 713 exists 12 and CONTAINS 11 712 CONTAINS 12 13 and 1 2 CONTAINS SOUTHERN US TEXAS and 1 3 CONTAINS TEXAS DALLAS The above explanation tells us that a rule our transitivity rule was invoked during the proof and that two ground assertions CONTAINS SOUTHERN US TEXAS and CONTAINS TEXAS DALLAS were accessed to supply preconditions for the rule These combined as sertions lead to the conclusion CONTAINS SOUTHERN US DALLAS Within an explanation directly asserted propositions are indicated with the prefix We can also ask why after a retrieve query However if the query has multiple solutions each one has a separate explanation In order to ask why we need to ask for one solution at a time This can be done by omitting the word all from the retrieve query and subsequently calling retrieve to obtain results one at a time retrieve contains southern us x 1 X DALLAS 5 The order of solutions will not necessarily be the same as shown here Chapter 3 Annotated Example 21 retrieve gt There are 2 solutions so far 1 X DALLAS 2 X TEXAS retrieve gt There are 3 solutions so far 1 X
67. 1 m pk ERE RP 41 create enumerated list ss 42 create enumerated set esse 42 reate float vector alli ua ess 115 crea t f nctiOn i cseicueeseseae dies aed ate bee 42 Greate KeywOrd 143 pcieeteteess Coke eet ese 130 create marker storag essc p Rus 115 create mo dule i lilius Ree ERE 42 ereate object i bg ce hir v a 42 create relation cece cee eese 42 ereate symDOl ev E RYE Les 130 create vector cllavgea day bep Ria 115 current inference level 115 QUI e cocen ete cue dies cide haa EQNPQUF TAURI E 71 142 D default false ee RR EEEEIET Dias 115 defa lt true dli el etait oe 116 defa lt truth value 2 3 nos 116 defconcept e bb o Headed ete Ee ne RT IPS 30 deffunction sioe y 6g esstePELRR ue Ee 31 define arithmetic operation on wrappers bed bius MM eS ERI d beet entrar 116 define arithmetic test on wrappers 116 define computed constraint 116 defirnstance sedeer risigne eenn pii reae h 31 defmodules eee itv eb EPI red 31 defobject i hp ehe hires n pex PER Ree e 33 defproposition eveeee 9 neces eee IR 33 defrelatiOn iee mc De mmni iia sen RS 33 defr le sleep ADS ON EEE 34 delete rul68 icosdced rnirar niteisiin 34 Ip C 34 deny 22st nas bate Siga tee Mes peed eeesd 34 deobjectify tree oiciee gae o ea Eres 116 describe fiiseycd kg eda d RR thd a6 34 describ
68. 12 and 713 When PowerLoom parses a top level proposition it automatically supplies universal quantifiers for any unquantified variables So the above rule is equivalent to the rule defrule transitive contains forall 11 712 13 gt and contains 11 712 contains 12 7 13 contains 11 13 Note Instead of defining a transitive contains rule we could have achieved the same effect by asserting that the contains relation is transitive e g by stating assert transitive contains Now that we have told the system that our contains relation is transitive let us rerun our query Actually PowerLoom isn t quite as strict as just stated its search for an identical proposition can acco modate changes in the names of variables Chapter 3 Annotated Example 20 retrieve all contains southern us x There are 3 solutions 41 X TEXAS 42 X AUSTIN 43 X DALLAS 3 11 Explanation PowerLoom provides a command called why that you can use to get an explanation of the logic behind one of its answers The why command explains the last query entered into the system i e it should invoked after one has submitted a retrieve or an ask command Before asking a why command you must enable the justifications feature set feature justifications Queries execute a bit more slowly with jusifications enabled which is why it is disabled by default Having enabled justifications we must re run a query Here is how
69. 129 test type on instan6Ce lv voee ees 129 THIRD ELEMENT 22 1 pGG Goan eed cde eee de 82 time conmilandz eei x 4v DPI AREE Y SD Y 38 TTA uiui nsennicue reed nta Ec RELIER pedcs 82 TRANSITIVE erw sRRIESBe TG p RIED 82 translate loom file esce een 129 true truth v alue ipa e hebes 129 BUS Fe sss ise EUER tr b S RR UU ERREUR 129 TYPEHOP 2i age e e Rh RIERRPE E ERES 82 U UNASSSL DP 129 unassert proposition sese 129 unknown truth value esses 129 unknown cua ue xx a he eae e ee 129 unset featU re ciu esee ee Re acaeves 39 unset powerloom feature 129 upclassify all descriptions 129 upclassify all instances ss 130 upclassify instances eee 130 upclassify named descriptions 130 update proposition in store on OBJECT STORE rem 130 update tple ieenesrbiebeRDPG 9 eve 130 V VALUE eer he ERES ROSE ERR PRESS 82 VARIABLE ARITY one RR eR ER Res 82 VARIANCE snIeic et t Aoc RE PIE RIP 83 W weaken truth value sees 130 Why cnssssisnccuquida d kcu es eR PIG PSP RUDI 39 with logic environment 130 within classification session 130 within meta cache cies s te es 130 Variable Index Variable Index A abstract Of iwi sive td etter ide see tirria aia 19 all super contexts Of s eco ER RES 71 ny value Of 1ll Rl ee ones 70 75 80 arguments Ol iie eeRRp
70. 85 8 amp 2 GUI Design Goals eme ber eee eee a E D ira 86 8 3 GUL OVeEVIeW aas ne ea db ER IU usd R R e RR RE 88 8 4 GULFPea8btureszsi Made be ODDO EECRRERTRPIIAQ OPEP TA 90 8 4 1 Connect tO Server i o ceireb edere bete trie patat 90 8 4 2 Edit Preferences 2 2 cfs cases t Reese EE ARE eds 90 84 3 JCOBOLoad 52yBuor secius Urs epa LR D CHE eub xad id id 90 S44 BIOWSIDE ione me Ex I epa ees enia re abes 90 eub EdUHEGzetesse toS y seh ae bbe E RE ad OA REPE wees E 93 GLO CIo08ePscenseceeEPsd re E nada Goes eee anes 95 8 4 7 Extension Editor esee 96 8 4 8 Ask and Retrieve Queries 000 cee eee eee eee 97 EAU PCE Cl corer erect tence ted esu cet E denm deatete Eee 98 S410 Conso les iis e ERE AIDSORIDSERerdRNPEPTe tenes 99 SAL Cut Copy Paste Delete sctcccsnsecetentsesiatcecurasex 100 wb Pate Work euo e ee eret a aus erac REFERRE 100 85 1 arte KIBS is ence pex and urs ente Resta e Een 101 8 9 2 Undo oeste DUE hU had ena ORE se 101 8 5 3 Drag and Drop eere ore eR ERU p 101 SoA ioctapbOOku s esed seit a E EEA EE EEE 101 8 5 5 Instance Cloning 0 cece eect nee 101 8 5 0 SECUTI Y 25 sib cu ace sa E Ia PECIA s AUI tede cae bis 101 amp b T Multiple Users eerte RR Dee E IP EXE E 101 9 Ontosaurus Lom osi dre roi t cba d cera daa 103 10 Installation eese eR RE rex Rh 104 10 1 System Requirements 0 0 cece cece eee ne 104 10 2 Unpacking the Sources ena E ES Ea a
71. 9 PowerLoom Console 8 4 11 Cut Copy Paste Delete The PowerLoom GUI supports cut copy paste and delete operations These operations can be used to edit text and in some cases they can be used to edit objects in lists or trees For example the concept hierarchy can be edited within the browser by selecting a con cept executing a cut operation selecting another concept and then executing paste This sequence of operations will delete the concept from it s original position in the hierarchy and make it a subconcept of the concept that was selected when the paste operation was performed The GUI implements a robust data transfer framework that is capable of recognizing the types of objects that are being transferred and the types of potential transfer sources and destinations This allows the application to prohibit nonsensical data transfers such as cutting a concept in a concept navigation pane and then trying to paste it into a module pane It also allows data transfer operations to be context sensitive For example cutting a concept in a concept navigation pane means that a move operation is being initiated while cutting a concept in a concept editor s superconcept list means that the concept should be removed from the list Additionally copying an object such as a concept then executing a paste inside a text window will paste the name of the object As one would expect text may be cut copied and pasted between the GUI and out
72. E BOOLEAN Function Tests whether tv is a known truth value i e true or false is true proposition1 relation and arguments OBJECT Function module MODULE environment ENVIRONMENT BOOLEAN Return TRUE if a proposition relation args has been asserted or inferred by forward chaining load in module filename STRING module MODULE Function environment ENVIRONMENT Read logic commands from the file named filename and evaluate them If the file does not have an in module declaration that specifies the module within which all remaining commands are to be evaluated it will be loaded in the module specified If no module is specified and the file does not contain an in module declaration an error will be signaled The remaining commands are evaluated one by one applying the function evaluate to each of them load native stream in module stream NATIVE INPUT STREAM Function module MODULE environment ENVIRONMENT Read logic commands from the native input stream stream and evaluate them As sumes stream is a line buffered stream which is a safe compromise but does not generate the best efficiency for block buffered streams such as files If the stream does not supply an in module declaration that specifies the module within which all remaining commands are to be evaluated it will be loaded in the module specified If no module is specified and the file does not supply an in module declaration an error will be si
73. E BOOLEAN Function Return TRUE if self represents some form of truth true self PROPOSITION BOOLEAN Function Return true if self is true or default true if we are considering default assertions unassert proposition PARSE TREE OBJECT N Command Retract the truth falsity or inconsistency of proposition This is a more general version of retract that also handles falsity For example if we assert the proposition not sad Fred and then execute the statement unassert sad Fred the truth value of the proposition sad Fred will be set to UNKNOWN If we had called retract in place of unassert the proposition sad Fred would remain set to FALSE Note that for this unassertion to succeed the logic constant Fred and the relation sad must already be defined unassert proposition self PROPOSITION Function Retract the truth falsity or inconsistency of the proposition self unknown truth value self TRUTH VALUE BOOLEAN Function Return TRUE if self represents UNKNOWN unknown self PROPOSITION BOOLEAN Function Return true if the truth of self is unknown unset powerloom feature feature KEYWORD Function Disable the PowerLoom environment feature feature upclassify all descriptions Function Classify all named descriptions Chapter 11 Miscellaneous 130 upclassify all instances Function Classify all named instances upclassify instances module MODULE local B
74. ECTION gt 1 LIST Function Project elements in column i zero based of the tuples of c and collect them into a list l PROPER SUBRELATION r RELATION sub RELATION Relation True iff sub is a proper subrelation of r written in set notation sub r This relation will generate bindings for at most one unbound argument PROPER SUPERRELATION r RELATION super RELATION Relation True iff super is a proper superrelation of r written in set notation super gt r This relation will generate bindings for at most one unbound argument PROPOSITION CONTEXT SENSITIVE OBJECT DYNAMIC SLOTS MIXIN Class BACKLINKS MIXIN home context CONTEXT Slot of Not documented kind KEYWORD Slot of Not documented truth value TRUTH VALUE Slot of Not documented arguments VECTOR Slot of Not documented operator GENERALIZED SYMBOL Slot of Not documented dependent propositions NON PAGING INDEX OF PROPOSITION Slot of Not documented Chapter 7 Built In Relations 78 PROPOSITION ARGUMENT p PROPOSITION i INTEGER gt Function arg THING Return the i th arg of p zero based PROPOSITION ARGUMENTS p PROPOSITION gt args LIST Function Return all arguments of p as a list args PROPOSITION ARITY p PROPOSITION gt arity INTEGER Function Return the number of arguments in p PROPOSITION RELATION p PROPOSITION gt 0p RELATION Function
75. ER Relation Relation that returns the positive and negative square roots y sqrt x For positive roots only see function SQRT Chapter 7 Built In Relations 81 STANDARD DEVIATION COLLECTION sd NUMBER Relation Binds sd to the standard deviation of the numbers in 71 STRING CONCATENATE x THING y THING gt z STRING Function Concatenate x and zero or more strings y variable arity and bind z to the result Coerces any type argument to a string if necessary STRING MATCH pattern STRING object THING start INTEGER Function end INTEGER gt match position INTEGER Match pattern against object between start and end zero based and return the position of the first match or fail if no match exists Supplying 1 for end indicates the end of object object can be a named logic object or a string pattern will eventually allow support regular expressions currently it only handles string literals Apart from doing to string coercion on object this is somewhat redundant since substring can generate start end pairs if its string and substring arguments are bound STRING MATCH IGNORE CASE pattern STRING object THING Function start INTEGER end INTEGER gt match position INTEGER Match pattern against object between start and end zero based and return the position of the first match or fail if no match exists Supplying 1 for end indicates the end of obje
76. F 15 acM 130 P CARRIED BY 600G TANK F 15 P CARRIED BY AIM 120 F 15 P CARRIED BY AIM 7M F 15 designed to gain and maintain air superiority in aerial combat Module AIRCRAFT KB Propositions P ALL WEATHER F 15 P CARRIED BY 20MM VULCAN F 15 P CARRIED BY 600G TANK F 15 Retrieve all results Retrieve Timeout Inference Level NORMAL Moveout Show Advanced Option Continue Current Module AIRCRAFT KB NTH DOMAIN GUIDANCE 1 GUIDANCE TYPB will be copied if you do a paste command Figure 8 2 The PowerLoom GUI The menu bar contains the following menus e KB menu items for connecting to a server loading saving and clearing KBs and opening browser and console windows e Edit menu items for cutting copying pasting and deleting and also an item which opens a preferences dialog e Objects menu items for opening editors on various KB objects including modules concepts relations instances and propositions This menu also contains an item to edit the currently selected object Chapter 8 PowerLoom GUI 89 e Query menu items for querying the KB searching the KB and editing a relation s extension e View this menu contains various items for updating the appearance of the application including a refresh item to bring the GUI up to date with the state of the KB on the server and
77. For a null item the column count is zero For non sequence objects the column count is one Chapter 6 PowerLoom API 44 get concept name STRING module MODULE Function environment ENVIRONMENT LOGIC OBJECT Return a class concept named name that is local to or visible from the module mod ule get concept instance matching value concept LOGIC OBJECT Function relation LOGIC OBJECT value OBJECT module MODULE environment ENVIRONMENT OBJECT Return a member of concept concept that has an attribute matching value for the binary relation relation i e relation result value holds get concept instances concept LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR Return instances of the concept concept Include instances of subconcepts of concept Depending on concept the return values could be wrapped literals get concept instances matching value concept LOGIC OBJECT Function relation LOGIC OBJECT value OBJECT module MODULE environment ENVIRONMENT PL ITERATOR Return members of concept concept that have an attribute matching value for the binary relation relation i e relation result value holds get current module environment ENVIRONMENT MODULE Function Return the currently set module get direct concept instances concept LOGIC OBJECT Function module MODULE environment ENVIRONMENT PL ITERATOR Return instances of concept concept Excl
78. G Function Return the fifth element e of FILLERS r RELATION i THING gt members SET Function Given a relation r and instance i returns a set of known fillers of r applied to i IMPORTANT this also collects intensional fillers such as skolems that might be identical extensionally FIRST ELEMENT l COLLECTION gt e THING Function Return the first element e of 71 FORK test PROPOSITION then PROPOSITION else PROPOSITION Relation Fail based conditional If test succeeds evaluates then otherwise evaluates else FOURTH ELEMENT l COLLECTION gt e THING Function Return the fourth element e of 71 FRAME PREDICATE c RELATION Concept A frame predicate is a second order relation that is used to describe constraints on the set of fillers for a binary relation applied to an instance Examples of frame predicates are range cardinality range type and numeric inclusive minimum Frame predicates are typically used to capture the kinds of relations manipulated by description logics such as USC ISI s Loom and W3C s OWL METHOD SLOT SLOT Class Not documented method setter BOOLEAN Slot of Not documented method parameter names LIST OF SYMBOL Slot of Not documented Chapter 7 Built In Relations 73 method parameter type specifiers LIST OF TYPE SPEC Slot of Not documented method return type specifiers LIST OF TYPE SPEC Slot of Not documented
79. IST n INTEGER gt r LIST Function Return the n th rest r of 7 zero based Count from end of the list if n is negative NUMERIC MAXIMUM r RELATION i THING n NUMBER Relation Relation that specifies an upper bound n on any numeric value that can belong to the set of fillers of the relation r applied to 71 NUMERIC MINIMUM r RELATION i THING n NUMBER Relation Relation that specifies a lower bound n on any numeric value that can belong to the set of fillers of the relation r applied to 71 Chapter 7 Built In Relations 77 NUMERIC SET s COLLECTION Concept s is a set of numbers OBJECT NAME x THING gt c STRING Function The name of the object X as a string This is just the name with module prefixes NOT included ORDERED c COLLECTION Relation 7c is ordered if the ordering of its members is significant Lists are ordered while sets are not PHRASE r THING 7s STRING Relation A phrase is a variablized sentence a template that is used to express individual ax iomatic facts as natural language sentences By convention a phrase contains one or more occurrences of each variable in a relation or concept definition it does not begin with a capital letter and it has no concluding period Systematic attachment of phrases to relations can be leveraged by tools that generate natural language para phrases of logic sentences PROJECT COLUMN i INTEGER c COLL
80. MED DESCRIPTION Function Return a set of all of the types that are asserted to be satisfied by self all class instances type SURROGATE CONS Function Return a set of instances that belong to the class type all cycles module MODULE local BOOLEAN CONS OF CONS Function Return a list of lists of descriptions that are provably co extensional all direct subrelations relation NAMED DESCRIPTION Function removeEquivalents BOOLEAN CONS OF NAMED DESCRIPTION Return a set of relations that immediately specialize relation If removeEquivalents recommended don t include any relations equivalent to relation all direct superrelations relation NAMED DESCRIPTION Function removeEquivalents BOOLEAN CONS OF NAMED DESCRIPTION Return a set of relations that immediately subsume relation If removeEquivalents recommended don t include any relations equivalent to relation all direct types self OBJECT CONS OF LOGIC OBJECT Function Return a set of most specific types that are satisfied by self Chapter 11 Miscellaneous 110 all equivalent relations relation NAMED DESCRIPTION Function reflexive BOOLEAN CONS OF NAMED DESCRIPTION Return a list of all relations equivalent to relation If reflexive include relation in the list all facts of instance self OBJECT Function includeunknownfacts BOOLEAN elaborate BOOLEAN LIST OF PROPOSITION Return a list of all definite TRUE
81. NG environment ENVIRONMENT MODULE Function Return a module named name get modules kb modules only BOOLEAN PL ITERATOR OF Function MODULE Return all modules currently loaded into PowerLoom If kb modules only is true then Stella modules that are used only for program code are not included in the list get name obj OBJECT STRING Function Return the fully qualified name of obj if it has one Otherwise return null get nth domain relation LOGIC OBJECT n INTEGER LOGIC OBJECT Function Return the type a concept for the the nth argument of the relation relation Count ing starts at zero NOTE if there are multiple nth domain propositions for relation this arbitrarily returns one of them it does not look for the most specific one which might have to be created get nth float sequence OBJECT n INTEGER module MODULE Function environment ENVIRONMENT FLOAT Return the floating point value in the nth column of sequence Counting starts at zero sequence must be of type proposition cons vector or PL iterator A zero column number returns a proposition s relational predicate For the PL iterator case the the current value pointed to by the iterator is used If this is not a floating point value then an exception will be thrown As a special case a column number of zero will also return the floating point value of sequence itself if it is not one of the types enumerated above This allows the u
82. OOLEAN Function Classify instances local to module and inherited by module If local don t classify inherited descriptions If module is NULL classify descriptions in all modules upclassify named descriptions module MODULE Function local BOOLEAN Classify named descriptions local to module and inherited by module If local don t classify inherited descriptions If module is NULL classify descriptions in all modules update proposition in store store Method on OBJECT STORE proposition PROPOSITION update mode KEYWORD A module with store has had the truth value of proposition change according to update mode The default method does nothing update tuple relation SURROGATE arguments CONS OF OBJECT Function updatemode KEYWORD PROPOSITION Assert or retract a proposition that applies relation to arguments weaken truth value tvl TRUTH VALUE tv2 TRUTH VALUE Function TRUTH VALUE If tv2 has lesser strength than tv1 adapt the strength of tv1 not its value and return the result Otherwise return tv1 unmodified with logic environment moduleForm OBJECT environment OBJECT Macro amp body body CONS OBJECT Execute body within the module resulting from moduleForm module is an ac ceptable moduleForm It will locally rebind module and context and shield the outer bindings from changes within classification session descriptionorinstance KEYWORD Macro amp body body CON
83. PowerLoom Initializing STELLA Initializing PowerLoom Loading required system webtools Loading required system ontosaurus Welcome to PowerLoom 4 0 0 Copyright C USC Information Sciences Institute 1997 2010 PowerLoom is a trademark of the University of Southern California PowerLoom comes with ABSOLUTELY NO WARRANTY Type copyright for detailed copyright information Type help for a list of available commands Type demo for a list of example applications Type bye exit halt quit or stop to exit PL USER Alternatively PowerLoom can be started in its standard command loop mode and then the GUI can be started with the start powerloom gui command This is useful if the Chapter 8 PowerLoom GUI 85 GUI becomes necessary at some point during development or if a different port should be used For example 4 powerloom Initializing STELLA Initializing PowerLoom Welcome to PowerLoom 3 2 52 Copyright C USC Information Sciences Institute 1997 2010 PowerLoom is a trademark of the University of Southern California PowerLoom comes with ABSOLUTELY NO WARRANTY Type copyright for detailed copyright information Type help for a list of available commands Type demo for a list of example applications Type bye exit halt quit or stop to exit PL USER start powerloom gui port 9092 Loading requir
84. PowerLoom Manual Powerful knowledge representation and reasoning with delivery in Common Lisp Java and C Version 1 48 16 October 2010 This manual describes PowerLoom 3 2 or later The PowerLoom development team Hans Chalupsky Robert M MacGregor Thomas Russ fhans tar isi edu Copyright 2010 University of Southern California Information Sciences Institute 4676 Admiralty Way Marina Del Rey CA 90292 USA Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies PowerLoom is a registered trademark of the University of Southern California Table of Contents 1 Jn PodOcblOE uae oe har rd rho obe acl rd 1 1 1 Powerloom Features 00 ccc ccc cece nn 1 1 2 Powerloom LHistOrys iu II d e 9 Ra n ctu tere eds 4 1 3 Running PowerLoOm i nee Pen tin esata ade xe 4 1 3 1 Command Line Options sseseseeeeeeeeeeeee 5 2 Conceptual Framework 6 2 1 Terms and Propositions ssseeseseseeeee sees 6 2 2 Dennitions 242 5 lt 1c 0 F200 be p e hRRIPDETNMRPbe e easy PRIMERS T 2 9 r thr Valuegaaaaut ensi eder E E E E ee aes 8 24 Modules RERRRRPEP ts EE GGG EEXERRRN ANGE eed RR RUE RS 9 3 Annotated Example 11 3 1 Using Modules 4 onere aee tates bee debe des 11 3 2 Concepts escrieu eraa n n Os n p ee RCRUM EE MAS 11 Deo Relations ernea tte pss
85. S OBJECT Used during classification Execute body within the indicated classification session and inference world within meta cache amp body body coNs OBJECT Macro Execute body within the meta cache of the current module Set appropriate special variables create keyword name STRING KEYWORD Function Returns the Stella keyword name creating it if necessary name is treated case sensitively This should generally not be necessary to do create symbol name STRING module MODULE Function environment ENVIRONMENT SYMBOL Returns the Stella symbol name visible in module creating it if necessary name is ALWAYS treated case sensitively even if module is case insensitive This should generally not be necessary to do Chapter 11 Miscellaneous 131 get keyword name STRING KEYWORD Function Returns the Stella KEYWORD name if it exists Case sensitive get name in module obj OBJECT module MODULE Function environment ENVIRONMENT STRING Return the name qualified as necessary so that obj can be found from module If there is no name for the object return null get short name obj OBJECT STRING Function Return the short name of obj if it has one Otherwise return null get symbol name STRING module MODULE Function environment ENVIRONMENT SYMBOL Returns the Stella SYMBOL name visible in module if it exists name is ALWAYS treated case sensitively is known tv TRUTH VALU
86. SER start powerloom server Chapter 8 PowerLoom GUI 86 Loading required system webtools Loading required system ontosaurus PL USER The command also accepts a port option to make the server listen to a different port Alternatively the PowerLoom Web Archive available in powerloom x y z native java lib ploom war can be deployed in an Apache Tomcat server by copying it into the server installation s webapps directory The GUI can then connect to it at the server port used by Tomcat 8 2 GUI Design Goals The choice of technology was driven by a number of design goals which are summarized below Note that the most of the GUI code was developed back in 2002 and is therefore somewhat outdated with respect to today s Java and Web based human interface technolo gies Visibility Knowledge Bases are complex and loosely structured entities It is often desirable to simultaneously maintain multiple views of a KB and to simultaneously perform multiple complementary task such as browsing editing querying and searching a KB With Swing s MDI Multiple Document Interface mode many internal frames can be open at the same time within a single desktop frame Swing also offers a very rich set of components and UI mechanisms which facilitate efficient use of screen real estate We designed the GUI to take advantage of Swing s presentation strengths We used the MDI mode so browsers editors etc can coexist on the same desktop Addit
87. STELLA load path and return the removed ele ment presume proposition PARSE TREE OBJECT N Command Presume the default truth of proposition Return the presumed proposition object KIF example presume happy Fred states that Fred is most probably happy Note that for this to succeed the relation happy must already be defined see assert print features Command Print the currently enabled and available PowerLoom environment features print rules relation OBJECT N Command Print the list of true rules associated with relation process definitions Command Finish processing all definitions and assertions that have been evaluated loaded since that last call to process definitions PowerLoom defers complete processing of definitions to make it easier to support cyclic definitions Following finalization of definitions this call performs semantic validation of any assertions evaluated since the last call to process definitions PowerLoom calls this function internally before each query the primary reason to call it explicitly is to force the production of any diagnostic information that results from the processing and validation Chapter 5 Commands 36 propagate constraints amp rest name NAME N Command Trigger constraint propagation over all propositions of module name If no name is supplied the current module will be used This also enables incremental constraint propagation for future monotonic up
88. TANCE Figure 8 8 Search Dialog When the user executes the search by hitting RET or selecting the OK button a list of results is presented These results are provided in a table format where one column is the name of the retrieved object another column contains the module that the object resides in and the final column specifies the type of the object i e concept instance etc As is the case with query results clicking on a search result item will update the topmost browser to display the selected object 8 4 10 Console The console window as shown in Figure 8 9 can be opened by selecting the KB gt Open PowerLoom Console menu item or by typing Ctrl p This opens an internal window which allows PowerLoom commands to be typed directly and sent to the PowerLoom server The response generated by PowerLoom is sent back to the GUI and printed below the prompt This functionality is similar to that of a LISP listener Chapter 8 PowerLoom GUI 100 Powerloom Console gt retrieve x wingspan x Aj F 16 33 gt retrieve all x wingspan x Aa F 16 33 F 15 42 F 117 43 4 F 111 63 EF 111 63 E 3 131 CF 18 40 4 C 9 93 C 5 223 C 22 108 C 21 40 C 20 78 C 17 171 C 141 160 C 130 133 B 52 185 B 2 172 B 1 137 AIM 9M 2 AIM 7M 3 3 AIM 120 1 7 AGM 88 3 7 AGM 84 3 AGM 130 5 A 6 53 A 10 57 5 AIRCRAFT KB Cancel Figure 8
89. TELLA powerloom has to be called explicitly However in Lisp it is not really necessary to use the command interpreter since all commands can also be executed directly at the Lisp top level The interactive command interpreter functions as a simple read eval print loop that prompts for input with a prompt reads a user command from standard input evaluates it and prints the result to standard output To exit the command interpreter type quit or stop To evaluate commands directly from a program the PowerLoom API provides the fol lowing evaluator functions evaluate command OBJECT module MODULE Function environment ENVIRONMENT OBJECT Evaluate the command command within module and return the result Currently only the evaluation of possibly nested commands and global variables is supported Commands are simple to program in Common Lisp since they are built into the lig you are executing within a case sensitive module then you may see some differences in behavior between commands evaluated by the command interpreter and commands invoked from the Lisp Listener Chapter 4 Communicating with PowerLoom 26 language and relatively awkward in Java and C Users of either of those languages are more likely to want to call s evaluate evaluate string expression STRING OBJECT Function Evaluate the expression represented by expression and return the result This is equivalent to evaluate unstringify expre
90. TEXAS DALLAS and 1 2 3 3 12 Contexts and Modules The final feature that we will illustrate in this section is the PowerLoom context mechanism PowerLoom organizes its knowledge into a hierarchy of logical containers called contexts A PowerLoom context is either a module a somewhat heavyweight object that includes its own symbol table or a world a very lightweight object designed for fast switching from one world to another All contexts inherit from a single root context The most important feature of a context is that a fact asserted into it is inherited by all contexts below it However a parent context is unaware of any knowledge entered into one of its descendants Here we concern ourselves only with modules We first define a second module called alternate business to be a subcontext of our business module and then we switch into the new module defmodule ALTERNATE BUSINESS includes BUSINESS in module ALTERNATE BUSINESS Next within the scope of the alternate business module we will create a new com pany And just for good measure we will change the name of MegaSoft while we are at it assert and company web phantoms company name web phantoms Web Phantoms Inc retract company name megasoft MegaSoft Inc assert company name megasoft MegaZorch Inc First here are pairs of companies and company names from the vantage point of the Business module in module
91. X4 ERE RUORETRN FR We B base module of 0 ccc ccc cece eee 71 child contexts Of ssiccic scwcd es cada ehe 71 context number of 0 00 cece cece 71 D dependent propositions of rg F function code of 0 ccc cece eens 73 H home context of 0c ccc cece eee eens Ta 147 kind Of Q2 Ds T awenbotise fS dpascg gres T method code of e E IAR e ER Ra 73 mnethod function OL o ope e bee RISE 73 method parameter names of 72 method parameter type specifiers of 73 method return type specifiers of 73 method setter of sse eese T2 method stringified source of 73 operator ofi iae eR a pE ENEE ences 77 surrogate value inverse of 82 T the cons list of sescenG es saad weer anno es TO tr th value of l1 oki eo hy bedi eee 77 Concept Index 148 Concept Index Index is nonexistent
92. a concept for fillers of the binary relation relation get relation name STRING module MODULE Function environment ENVIRONMENT LOGIC OBJECT Return a concept or relation named name that is local to or visible from the module module get relation extension relation LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return propositions that satisfy relation Include propositions that satisfy subrela tions of relation get rules relation LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return rules attached to the concept relation relation in either antecedent or conse quent position Chapter 6 PowerLoom API 48 get types object LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF LOGIC OBJECT Return all named concepts that object belongs to initialize Function Initialize the PowerLoom logic system This function needs to be called by all appli cations before using PowerLoom If it is called more than once every call after the first one is a no op is a object OBJECT concept LOGIC OBJECT module MODULE Function environment ENVIRONMENT BOOLEAN Return TRUE if object is a member of the concept concept is default tv TRUTH VALUE BOOLEAN Function Tests whether tv is a default truth value is enumerated collection obj OBJECT BOOLEAN Function Test whether obj
93. a registration entry for name as a specialist which executes code Essentially just builds the Stella meta information tructure needed to funcall name as a spe cialist The function definition in code needs to accept a CONTROL FRAME and KEYWORD as arguments and return a KEYWORD Side effects on elements of the proposition in the control frame can be used to bind and thus return values register specialist function name stella name STRING Command native name STRING registers a specialist function stella name based on the native name for the particular programming language in question Use of this command makes the resulting code or knowledge bases non portable to other target languages relation name self NAMED DESCRIPTION STRING Function Given a relation object return it s name relation objectRef OBJECT BOOLEAN Function Return TRUE if objectRef denotes a relation or a class remove at self key LIKE ANY KEY SELF Method on QUERY SOLUTION TABLE Remove the solution identified by key from self To preserve the solution ordering chain the solution is marked as deleted and will be completely removed upon the next iteration through self remove deleted members self LIKE SELF Method on PAGING INDEX Destructively remove all deleted members of self Chapter 11 Miscellaneous 126 reset query caches Function Zero out all caches managed by the query optimizer so that it will reoptimize subgoal
94. age option to specify a list of package names separated by periods that prefix the Java object in this module Use java catchall class to specify the name of the Java class to contain all global amp special variables parameter less functions and functions defined on arguments that are not classes in the current module The default value will be the name of the module Chapter 5 Commands 33 When set to TRUE the api option tells the PowerLoom User Manual generator that all functions defined in this module should be included in the API section Ad ditionally the Java translator makes all API functions synchronized defobject amp rest args PARSE TREE LOGIC OBJECT N Command Define or redefine a logic instance The accepted syntax is defobject lt constant gt documentation lt string gt lt keyword option gt lt keyword option gt represents a keyword followed by a value that states an assertion about lt constant gt See defrelation for a description of lt keyword option gt s defobject provides a sugar coated way to assert a collection of facts about a logic constant but otherwise adds nothing in terms of functionality defproposition amp rest args PARSE TREE PROPOSITION N Command Define or redefine a named proposition The accepted syntax is defproposition lt name gt lt sentence gt documentation lt string gt forward only true false backward only true false don
95. al only return inconsistent propositions conceived locally in module all instances module MODULE local BOOLEAN ITERATOR OF Function LOGIC OBJECT Iterate over all instances or individuals visible from module Only instances that haven t been deleted will be considered If local only return instances created locally in module all named descriptions module MODULE local BOOLEAN Function ITERATOR OF NAMED DESCRIPTION Iterate over all named descriptions visible from module If local return only named descriptions interned in module If module is null return all named descriptions interned everywhere all named instances module MODULE local BOOLEAN Function ITERATOR OF LOGIC OBJECT Iterate over all named instances or individuals visible from module Only instances that haven t been deleted will be considered If local only return instances created locally in module Chapter 11 Miscellaneous 111 all named terms module MODULE local BOOLEAN ITERATOR OF Function OBJECT Iterate over all named terms visible from module A term can be an instance or individual as well as a description Only terms that haven t been deleted will be considered If local only return terms created locally in module all propositions module MODULE local BOOLEAN ITERATOR Function OF PROPOSITION Iterate over all conceived propositions visible from module Only propositions that haven t bee
96. al facts are not asserted globally but instead hold only within a specific context For example a logical proposition may evaluate as true within one module and evaluate as false within a different one The discussion below uses some examples of actual PowerLoom syntax to illustrate certain points However we gloss over the fine points of syntax and instead focus on semantic issues The next chapter reverses that emphasis and presents a series of examples that illustrate the breadth of syntactic constructs implemented for the PowerLoom language 2 1 Terms and Propositions A knowledge base attempts to capture in abstract machine interpretable form a useful representation of a physical or virtual world The entities in that world are modeled in the knowledge base by objects we call terms Examples of terms are Georgia denoting the U S state BenjaminFranklin denoting the historical person by that name the number three the string abc and the concept Person Unlike objects in an object oriented programming language the terms in a PowerLoom knowledge base usually have distinct names unless there are sufficiently many that naming them all becomes impractical Terms are categorized or related to one another by objects called relations Examples of relations are has age greater than is married to plus Concepts such as Person State Company and Number are considered a subcategor
97. ancestor modules If module is NULL classify relations in all modules Conceptually the classifier operates by comparing each concept or relation with all other concepts relations searching for a proof that a subsumption relation exists be tween each pair Whenever a new subsumption relation is discovered the classifier adds an implication link between members of the pair thereby augmenting the structure of the concept or relation hierarchy The implemented classification algo rithm is relatively efficient it works hard at limiting the number of concepts or relations that need to be checked for possible subsumption relationships Chapter 5 Commands 30 classify instances module NAME local BOOLEAN N Command Classify instances visible in module If local only classify instances that belong to module i e don t classify instances inherited from ancestor modules If module is NULL classify instances in all modules Conceptually the classifier operates by comparing each instance with all concepts in the hierarchy searching for a proof for each pairing indicating that the instance belongs to the concept Whenever a new is a relation is discovered the classifier adds an is a link between the instance and the concept thereby recording an additional fact about the instance The implemented classification algorithm is relatively efficient it works hard at limiting the number of concepts or relations that need to be checked
98. arate line on the standard output stream print goal stack frame CONTROL FRAME verbose BOOLEAN Function Print stack of goals Assumes that query has been interrupted with a full stack of control frames print logical form form OBJECT stream OUTPUT STREAM Function Print the logical form form to stream according to the current setting of logic dialect Pretty printing is controlled by the current setting of prettyPrintLogicalForms print logical form in dialect self OBJECT dialect KEYWORD Function stream OUTPUT STREAM Produce a stringified version of a logical representation of self and write it to the stream stream Use the dialect dialect or use the current dialect if dialect is NULL print unformatted logical form form OBJECT Function stream OUTPUT STREAM Print the logical form form to stream according to the current setting of logic dialect Pretty printing is explicitly forced to be turned off print whynot justification justification JUSTIFICATION Function stream OUTPUT STREAM maxDepth INTEGER style KEYWORD summary BOOLEAN Print WHYNOT justification to stream according to maxDepth and style Print a summary only if summary is TRUE random float n FLOAT FLOAT Function Generate a random integer in the interval 0 n 1 n must be lt 2715 recall marked objects self LIST ITERATOR Method on MARKER TABLE Return an iterator that generates all marked
99. asserted proposition assert nary proposition relation and arguments OBJECT Function module MODULE environment ENVIRONMENT PROPOSITION Assert that the proposition represented by the list relation and arguments satisfies the relation relation assert proposition proposition PROPOSITION module MODULE Function environment ENVIRONMENT PROPOSITION Assert that the proposition proposition is true in module Return the asserted propo sition assert unary proposition relation LOGIC OBJECT arg OBJECT Function module MODULE environment ENVIRONMENT PROPOSITION Assert that the proposition relation arg is TRUE in module Return the asserted proposition change module module MODULE MODULE Function Set the current module to module and return it If module is null then no switch is performed and the current module is returned clear caches Function Clear all query and memoization caches clear module module MODULE MODULE Function Destroy the contents of the module module as well as the contents of all of its children recursively conceive sentence OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Create one or more proposition objects from the sentence sentence in the module module Return an iterator of the propositions If any of the new propositions has the same structure as an already existing proposition an automatic check for duplicates
100. at there exists an explicit assertion of a proposition Q x1 xn and either Q equals P or Q is a proper subrelation of P In otherwords the only rules that imply membership in P are those that relate P to one of its proper subconcepts or subrelations Query A query probes the informational state of a knowledge base An ask query test the truth of its propositional argument A retrieve asks for sets of constants bindings that make its propositional argument true when the constants are substituted in place of its variables The propositional argument to ask and retrieve arbitrary expression in the first order predicate calculus Because of constraints imposed either by resource limitations or inherent undecidability PowerLoom cannot guarantee the completeness of its inferences Relation Retraction A retraction changes the truth value of a proposition from either True or False to the value Unknown Retraction is a procedural non declarative operation Rule A rule is any universally quantified proposition i e a proposition of the form forall x1 xn logical sentence with free variables x1 xn Power Loom supports several different syntactic constructs for defining rules See also Forward Rule and Backward Rule Subsumption A subsumption relation specifies the relative generality of two concepts A concept A subsumes a concept B if the definitions of A and B logically imply that members of B must also be
101. ava constructors should not be called directly Instead there will be a static method new lt ClassName gt with the class name in mixed case that should be used instead Most of the functions of interest will be in the edu isi stella Stella edu isi powerloom PLI or edu isi powerloom logic Logic classes Methods typically refer to their first argument as self Methods which return more than one return value will take a final argument which is an array of Stella_Object which will be used to return the additional arguments Primitive types in Stella have the following mapping in Java Stella Java INTEGER int FLOAT double NUMBER double CHARACTER char BOOLEAN boolean STRING String MUTABLE STRING StringBuffer NATIVE OUTPUT STREAM java io PrintStream NATIVE INPUT STREAM java io PushbackInputStream May change Inside Stella PowerLoom objects and collections primitive types are wrapped using Stella wrappers instead of Java s primitive classes So integers will be wrapped as edu isi stella Integer Wrapper rather than java lang Integer Wrappers have a field called wrapperValue which accesses the internal value Example of use import edu isi stella import edu isi stella javalib IntegerWrapper iWrap IntegerWrapper wrapInteger 42 int answer iWrap wrapperValue 6 2 3 3 Using the PLI Class To make interoperability between PowerLoom and Java a little simpler we are providing a PowerLoom Interface class name
102. ay not match name exactly Return the object create relation name STRING arity INTEGER module MODULE Function environment ENVIRONMENT LOGIC OBJECT Create a relation named name with arity arity in the designated module Domain and range information can be added via assertions of nth domain or domain and range relations Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly Chapter 6 PowerLoom API 43 destroy object object OBJECT Function Delete the object object retracting all facts attached to it empty self BOOLEAN Method on PL ITERATOR Return TRUE if the iterator self has no more elements evaluate command OBJECT module MODULE Function environment ENVIRONMENT OBJECT Evaluate the command command within module and return the result Currently only the evaluation of possibly nested commands and global variables is supported Commands are simple to program in Common Lisp since they are built into the language and relatively awkward in Java and C Users of either of those languages are more likely to want to call s evaluate generate unique name prefix STRING module MODULE Function environment ENVIRONMENT STRING Generates a name based on prefix with a number appended that is not currently in use in module In a non case sensitive module the returned name will be all upper case This latte
103. bject will be of type THING A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly Return the object s create module name STRING parent name STRING Function case sensitive BOOLEAN environment ENVIRONMENT MODULE Creates a new module name as a child of parent name The flag case sensitive controls whether names read in this module will be case sensitive or not s create relation name STRING arity INTEGER Function module name STRING environment ENVIRONMENT LOGIC OBJECT Create a relation named name with arity arity in the designated module Domain and range information can be added via assertions of nth domain or domain and range relations A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly Chapter 6 PowerLoom API 5A s destroy object object name STRING module name STRING Function environment ENVIRONMENT Delete the object named object name retracting all facts attached t
104. bles referenced in proposition will be used as the query variables If proposition is omitted the most recently asked query will be continued A solution is a set of bindings for the listed variables for which proposition is true The optional first argument controls how many solutions should be generated before control is returned The keyword a11 indicates that all solutions should be generated By default retrieve returns after it has found one new solution or if it cannot find any more solutions retrieve returns an iterator which saves all the necessary state of a query and stores all generated solutions When used interactively the returned iterator will print out with the set of solutions collected so far Calling retrieve without any arguments or only with the first argument will generate one or more solutions to the most recently asked query KIF examples retrieve happy x will try to find one happy entity and store it in the returned query iterator retrieve 10 happy x will try to find 10 happy entities retrieve 10 will try to find the next 10 happy entities retrieve all happy x will find all happy entities retrieve all x Person happy x will to find all happy people Here we used the optional retrieve variable syntax to restrict the acceptable solutions The above is equivalent to the following query retrieve all and Person x happy x Similarly retrieve all x Person
105. but they are mistaken Chapter 12 Glossary 135 Description A description is an expression that defines a particular logical relation e g the class of all three legged black cats In PowerLoom the terms concept and relation generally refer to named relations while a description may or may not have a name The KIF operators kappa and setofall are used to define unnamed descriptions Description Logic The term description logic refers to a logic that focuses on descrip tions as its principal means for expressing logical expressions A description logic system emphasises the use of classification and subsumption reasoning as its primary mode of in ference Loom and Classic were two early examples of knowledge representation systems that implement description logics Domain Model A collection of definitions rules and facts that characterizes the possible states of some real or imagined world The domain model specifies a terminology of concepts and relations that is useful for describing objects in that world Often domain model refers to that portion of a world s representation that does not change over time Extension Given a relation R with arity N the extension of R is the set of ground propositions of the form R x1 xN whose truth value is true If R is a concept then its extension of often considered to be not a set of unary tuples but the set of argument fillers of those tuples i e t
106. conceivable that the GUI code could be used as the basis for a more specific Chapter 8 PowerLoom GUI 88 application which would have its own application specific menus and windows but would retain the general purpose browsing querying and editing tools for direct manipulation of the knowledge base 8 3 GUI Overview An example screen shot of the PowerLoom GUI is shown in Figure 8 2 The main application frame consists of pull down menus a toolbar and a status bar KB Edit Objects Query View Navigate Window Help amp amp a amp l e e Browsing Knowledge Base blackcatisi edu 800 Concepts for AIRCRAFT KB amp f Relations for AIRCRAFT KB Il H ROOT MODULE H POWERLOOM SERVER 9 H PL KERNEL KB H CYC FRAME ONTOL H KIF FRAME ONTOLO H LOOM API Q9 Hn PL USER H AIRCRAFT KB CONCEPT ROOT ACTIVE GUIDANCE TYPE ENTITY c PHYSICAL ENTITY c SOCIAL ENTITY GROUND UNIT GUIDANCE TYPE t t t t t Query wingspan X Y RELATION ROOT ATTACKED BY CALIBER CARRIED BY CEILING COST CREW ROLES 1 Editing Instance F 1 F 111 F 117 F 118 GE 100 F 15 F 16 F100 PW 100 F110 GE 200 F 15 Documentation C ALL WEATHER Module AIRCRAFT KE The F 15 Eagle is an all weather extremely maneuverable tactical fighter FIXED WING AIRCRAFT Results Propositions for F 15 X P ALL WEATHER F 15 AR P CARRIED BY 20MM VULCAN
107. contexts scenar ios and hypothetical reasoning non monotonic overrides in sub contexts etc Pow erLoom s context mechanism is built in at a very low level using a very efficient and light weight implementation for maximum performance e Open and closed world By default PowerLoom makes an open world assumption and returns unknown if it cannot prove or disprove a question However concepts and relations can be selectively marked as closed to support selective closed world reasoning PowerLoom also has a fail predicate in addition to true negation via not to implement closed world negation as failure which can be useful in certain situations e Knowledge base management PowerLoom supports incremental monotonic and non monotonic updates that extend or non monotonically change the assertion base In PowerLoom one can effectively interleave definitions re definitions assertions and re Chapter 1 Introduction 3 tractions with retrieval and inference without having to reload large knowledge bases from scratch after every change Truth maintenance of cached inference results that might have been invalidated by updates is done via inference cache contexts After a knowledge base has been loaded and changed by some updates the changed state can be saved out to a file or an experimental persistent store built on top of a relational database e Scalability Despite its emphasis on expressive representation which usually has to be paid for w
108. copy self LIKE SELF Method on JUSTIFICATION Return a copy of the proof starting at self Allocates all new justification objects but structure shares other information such as propositions and substitutions create name GENERALIZED SYMBOL N Command amp rest type GENERALIZED SYMBOL OBJECT Create a logic object with name name and return it If type is also supplied assert that the object belongs to that type create 2 d array nof rows INTEGER nof columns INTEGER Function amp rest values OBJECT 2 D ARRAY Create a two dimensional array with nof rows rows and nof columns columns and initialize it in row major order from values Missing values will be padded with NULL extraneous values will be ignored create 2 d float array nof rows INTEGER nof columns INTEGER Function amp rest values FLOAT 2 D FLOAT ARRAY Create a two dimensional array with nof rows rows and nof columns columns and initialize it in row major order from values Missing values will be padded with NULL extraneous values will be ignored create float vector amp rest values FLOAT FLOAT VECTOR Function Return a vector containing values in order create marker storage supportRecall BOOLEAN MARKER TABLE Function Return a new marker storage object used to remember with objects have been marked lf supportRecall is set then the iterator recall marked objects can be invoked on the new marker storage object create ve
109. cription in module or in any module if module is NULL that were not or will not be classified due to their lack of non inferable primitive ancestor relations list undefined relations module NAME local BOOLEAN N Command CONS OF NAMED DESCRIPTION Return a list of as yet undefined concepts and relations in module These relations were defined by the system since they were referenced but have not yet been defined by the user If module is NULL look in the current module If local only look in module but not in any modules it inherits Chapter 11 Miscellaneous 122 listify self LIST Method on QUERY ITERATOR Just like QUERY ITERATOR consify but return a LIST instead load cmd line files Command Loads all PowerLoom files specified on the command line If directories are listed all PowerLoom files in those directories are loaded Since when this is called we might still have unprocessed command line args this only looks at files which are to the right of the last argument that starts with a character load directory directory STRING Command Load all PowerLoom files plm in directory in alphabetic sort order load stream stream INPUT STREAM Function Read logic commands from stream and evaluate them load stream in module stream INPUT STREAM Function default module MODULE Read logic commands from stream and evaluate them If default module is not null then any commands will be read in
110. ct object can be a named logic object or a string This match compares the strings ignoring differences in letter case SUBRELATION r RELATION sub RELATION Relation True iff sub is a subrelation of r written in set notation sub lt r This relation will generate bindings for at most one unbound argument SUBSET OF sub COLLECTION super COLLECTION Relation True if sub is a subset of super For performance reasons the subset of predicate refuses to search for bindings if both of its variables are unbound Implementation note subset of is treated specially internally to PowerLoom and hence Powerloom does not permit the augmentation of subset of with additional inference rules In otherwords subset of behaves semantically like an operator instead of a relation SUBSTRING s STRING start INTEGER end INTEGER gt Function sub STRING Generate the substring of s starting at position start zero based ending just before position end and bind sub to the result This is the PowerLoom equivalent to the STELLA method subsequence In addition this function can be used to locate substrings in strings by supplying values for s and sub and allowing start and end to be bound by the function specialist In other words retrieve all start end substring foo start end o gt start 1 end 2 start 2 end 3 SUBSUMPTION QUERY prop PROPOSITION Relation Query prop with infe
111. ctor amp rest values OBJECT VECTOR Function Return a vector containing values in order current inference level NORMAL INFERENCE LEVEL Command Return the current inference level that is active in the current query the current module or otherwise globally default false self PROPOSITION BOOLEAN Function Return true if self is default false Chapter 11 Miscellaneous 116 default true self PROPOSITION BOOLEAN Function Return true if self is default true default truth value self TRUTH VALUE BOOLEAN Function Return TRUE if self is a default truth value define arithmetic operation on wrappers name SYMBOL Macro operation name SYMBOL OBJECT Defines name as an arithmetic comparision operation using the test test name It will take two wrapped number parameters and return a wrapped number The code will use the appropriate test for the specific subtype of wrapped number actually passed in and return the appropriate subtype of wrapped number based on the normal arithmetic contagion rules For example if both input parameters are wrapped integers then the output will be a wrapped integer If the inputs are a wrapped integer and a wrapped float then the output will be a wrapped float etc define arithmetic test on wrappers name SYMBOL Macro test name SYMBOL OBJECT Defines name as an arithmetic comparision operation using the test test name It will take two wrapped number parame
112. d PLI which handles synchronization issues setting and restoring the reasoning context and the a more convenient use of some Java native objects rather than Stella objects Generally that means that strings are used for PowerLoom expressions and return values rather than Stella Cons objects Details about the methods can be found in the section Chapter 6 PowerLoom API page 40 The names of functions in that section will need to be converted to their Java Chapter 6 PowerLoom API 63 equivalents using the conventions described in Section 6 2 3 2 PowerLoom Java Conven tions page 60 We also provide javadoc documentation for the edu isi powerloom PLI class We recommend using this method for accessing PowerLoom functionality We expect to expand the range of PowerLoom interface functions that have an analog in the PLI class over time 6 2 3 4 Using Stella Objects Stella objects can also be used directly The most common ones used by PowerLoom users are Module and LogicObject Other potentially useful Stella objects are Cons Symbol Keyword and List Except for LogicObject these are in the edu isi stella package LogicObject is in the edu isi powerloom logic package If one wishes to construct Cons objects for example to create objects to pass to interface functions one would begin by building items up using Stella_Object cons static method which takes a stella object and a cons The empty cons is kept in the edu isi stella Stella NIL
113. dates to the module Once a non monotonic update is performed i e a retraction or clipping of a function value all cached inferences will be discarded and constraint propagation will be turned off until this function is called again push load path path STRING CONS OF STRING WRAPPER Command Add the directories listed in the separated path to the front of the STELLA load path Return the resulting load path repropagate constraints amp rest name NAME N Command Force non incremental constraint propagation over all propositions of module name If no name is supplied the current module will be used This also enables incre mental constraint propagation for future monotonic updates to the module similar to propagate constraints reset features LIST OF KEYWORD Command Reset the PowerLoom environment features to their default settings reset powerloom Command Reset PowerLoom to its initial state CAUTION This will destroy all loaded knowl edge bases and might break other loaded STELLA systems if they do reference Pow erLoom symbols in their code retract proposition PARSE TREE OBJECT N Command Retract the truth of proposition Return the retracted proposition object KIF example retract happy Fred retracts that Fred is happy Note that for this assertion to succeed the relation happy must already be defined If the constant Fred has not yet been created it is automatically created as a sid
114. defines the KIF syntax and semantics PowerLoom adopts KIF as its representation language and adds a few extensions Knowledge Base A knowledge base attempts to capture in abstract machine interpretable form a useful representation of a physical or virtual world The entities in that world are modeled in the knowledge base by objects we call terms Examples of terms are Georgia denoting the U S state BenjaminFranklin denoting the historical person by that name the number three the string abc and the concept Person Literal A logically static constant Examples are numbers strings quantities and truth values Chapter 12 Glossary 136 Module Open World Semantics PowerLoom assumes an open world semantics unless a user ex plicitly specifies that it use closed world semantics Under this assumption if PowerLoom cannot prove or disprove a proposition then it assigns that proposition the value Unknown See also Closed World Semantics Predicate The term predicate is a syntactic notion that refers to the zeroth arguments of a proposition Predicates denote relations and properties i e sets Proposition A logical sentence whose truth value can be evaluated with respect to some context Each PowerLoom assertion assigns the value True or False to some proposition Primitive Relation P is a primitive concept or relation if and only if a proof that P x1 xn is true exists only for the case th
115. dit Extensiun fur GASOLINE ENGINE FUEL TANK gt t MOVING OBJEC X Cut Ctrl X NAVAI IINIT B Cupy suse Paste Ctrl V E Delete Delete Figure 8 1 Context specific menu of possible actions on a concept Editability Applications that support text editing often need capabilities above and beyond the baseline copy cut and paste capabilities that all text widgets support In particular applications that allow editing of text with a regular structure such as source code or Lisp expressions may take advantage of special key bindings which augment basic navigation and editing capabilities The PowerLoom GUI makes use of Swing s powerful text components by implementing a full set of Emacs style keybindings These keybindings allow a user to perform such opera tions as navigating up and down a subexpression hierarchy selecting entire subexpressions and completing incomplete symbols In addition matching parenthesis are automatically highlighted in the GUI s text components Extensibility While it is not easy to claim that Swing applications are inherently more extensible than Web applications Swing s MDI architecture and pull down menu framework allows new features to be added with little disruption to the rest of the application In particular using modern GUI design tools that are part of IDE s such as Netbeans or Eclipse it is fairly easy to add new components and functionality to the PowerLoom GUI It is also
116. e concepts MAN and WOMAN are disjoint NAME TO OBJECT n THING gt o THING Function Find or create the PowerLoom logic object o named by the name n in the current module We are interpreting the name n literally here i e it is not a print name as assumed by PLI functions and an object with exactly that name will be looked up or created If n is not a string this will coerce it to a string first We are intentionally using a separate function here instead of folding this into object name since we might want to be able to look for an object with a certain name but not create one if it doesn t exist NORMAL QUERY prop PROPOSITION Relation Query prop with inference level set to normal Equivalent to query prop inference level normal but more efficient NTH DOMAIN r RELATION i INTEGER d COLLECTION Relation True if the nth value for a tuple T satisfying r must belong to the concept d Argument counting starts at zero NTH ELEMENT COLLECTION n INTEGER gt e THING Function Return the n th element e of 1 zero based Count from end of the list if n is negative If n is unbound and e is bound this computes the position of e If both are unbound collection elements and their respective positions will be enumerated NTH HEAD l LIST n INTEGER gt h LIST Function Return the n head elements h of 71 Count from end of the list if n is negative NTH REST l L
117. e effect of calling retract retract facts of instanceRef OBJECT N Command Retract all definite TRUE or FALSE propositions that reference the instance in stanceRef retract from query query CONS amp rest options OBJECT N Command CONS OF PROPOSITION Evaluate query which has to be a strict or partial retrieval command instantiate the query proposition for each generated solution and retract the resulting propositions See assert from query for available command options retract rule ruleName NAME PROPOSITION N Command If it is currently TRUE set the truth value of the rule named ruleName to UN KNOWN This command may be used alternately with assert rule to observe the effects of querying with or without a particular named rule being asserted within the current context The proposition having the name ruleName may be any arbitrary proposition although we expect that it is probably a material implication Chapter 5 Commands 37 retrieve amp rest query PARSE TREE QUERY ITERATOR N Command Retrieve elements of a relation tuples that satisfy a proposition The accepted syntax is retrieve lt integer gt all lt vardecl gt lt vardecl gt Xproposition The variables and proposition are similar to an exists sentence or kappa term without the explicit quantifier If variables are declared they must match the free variables referenced by proposition Otherwise the free varia
118. e no effect If a proposition is asserted to be true and subsequently is asserted to be false or vice versa a clash or contradiction results When a clash is detected by PowerLoom a clash exception is thrown The system s default behavior is for the exception to be caught and ignored with the result that an assertion that would otherwise cause a clash never takes effect Applications that execute commands slightly below the top level i e below the clash exception catcher can catch the exception themselves and perform a specialized response PowerLoom s proof by contradiction specialist catches clashes to determine that a contradiction has occurred If a user or application wants to assign a proposition a truth value that isn t stronger than the current value it must first retract the current value The PowerLoom retract operator has the effect of undoing a prior assertion For example if we assert that Mary is a parent of Fred and then retract that assertion the value of the proposition parent of Mary Fred becomes unknown The proposition can then be assigned any other truth value We should note that executing a retraction does not necessarily cause a proposition to cease being true Consider the following sequence defconcept Person Chapter 2 Conceptual Framework 9 defconcept Employee e gt Person e assert Person Mary assert Employee Mary retract Person Mary If we now ask PowerLoom whether or no
119. e object on NAMED DESCRIPTION 116 description name on DESCRIPTION 117 description name on NAMED DESCRIPTION 116 deStrOy c Rp RAIL E ud obbR Iure 34 destroy instanCe i dea be IR teeaee 117 destroy objeCt aceissssdasade ray er ers 43 117 destroy propoSition x ccc eR RR 117 desStroy term c eee tata de Meena 117 DIREGT SUBRELAT ION oi ci ced lee m ERR de 71 DIREGCI SUPERRELATION 2 e RR Re des 71 direct superrelatiOns see 117 disabled powerloom feature 117 disjoin truth value g 2ccc e da res 117 DISJOINT iue minder aene ate oe 71 DISJ INT COVERING entem mme Rem Tub disjoint t rms o2 EE Rer nen Dir do clear instanceS c n xg 117 do save module m re Re e yr RR 117 DOCUMENTATION nr eR T DOMAIN m RR Rr RR ane RARE HE RI 71 drop load patbh iee c eee ee PR 34 DUPLIGCATE PREE RR ne RARE S een 2 DUPLICATE FREE COLLECTION T2 E EMPTY 53 2 0 ei aad hat a lata a otk eed hells Scan ad 72 empty on FLOAT VECTOR 00 eee eee 117 empty on PL ITERATOR 0 cece eee eee 43 empty on QUERY SOLUTION TABLE 117 enabled powerloom feature 117 EQUIVALENT RELATION seeeeeeee T2 estimated length on PAGING INDEX TIT evaluate eve et man Mave eii ER Re 25 43 evaluate proposition sess 118 evaluate String zilosacec siga aa ads 26 evaluation
120. e user then selects the Navigate gt Back menu item the selection history will be rolled back so that only AIRCRAFT KB is selected If the user then selects Navigate gt Forward the selection history will be rolled forward to its original state so that both AIRCRAFT KB and GUIDANCE TYPE are selected Chapter 8 PowerLoom GUI 93 Navigation Knowledge can be explored by expanding and collapsing nodes in hierar chical navigation panes such as the concept and module navigation panes If a tree or list is not fully visible the user may use the scrollbar on the navigation pane s right hand side to scroll through the contents of the pane Detailed views of objects such as concepts and relations can be obtained by right clicking the object and selecting the Edit menu item To navigate to a constituent of a proposition the user can right click the constituent and then select the Navigate to menu item For example right clicking on the GUIDANCE argument in the proposition NTH DOMAIN GUIDANCE 1 GUIDANCE TYPE presents a popup menu which displays among other items the item Navigate to GUIDANCE Selecting this menu item will cause the browser to display and select the GUIDANCE relation Actions external to the browser may also update the browser s contents For example clicking on an instance in a list of query results will cause the browser to navigate to the selected instance Actions Right clicking inside the browser will present a menu o
121. ed ruleName Like get rule but find rule implicity quotes its input argument function relationRef OBJECT BOOLEAN Function Return TRUE if relationRef references a function generate expression self LOGIC OBJECT Function canonicalizevariablenames BOOLEAN OBJECT Return an s expression representing the source expression for self generate specialized term self OBJECT Method on LOGIC THING Method to generate a specialized term for self This is designed to allow for exten sion of the term generation code to cover other types of objects for the logic This particular method will signal an error unless there is a surrogate value inverse link set get class instanceRef OBJECT LOGIC OBJECT Function Return the nearest class with name instanceRef visible from the current module instanceRef can be a string symbol or surrogate If instanceRef is a surrogate the search originates in the module the surrogate was interned in get forward justifications proposition PROPOSITION LIST OF Function JUSTIFICATION Return propositions forward justifications get instance instanceRef OBJECT OBJECT Function Return the nearest instance with name instanceRef visible from the current module instanceRef can be a string symbol or surrogate If instanceRef is a surrogate the search originates in the module the surrogate was interned in get module moduleRef OBJECT MODULE Function Return a module named
122. ed system webtools Loading required system ontosaurus PL USER When the GUI is run against an embedded server as in the examples above it is run asynchronously and commands can be executed at the PowerLoom command loop at any time Moreover when the GUI is exited the command loop stays active until it is exited explicitly as well The GUI can also be run standalone in which case the user needs to connect to a running PowerLoom server For example powerloom gui only Running standalone PowerLoom GUI Once the GUI is initialized the user can connect to a server via the File gt Connect to Server menu item After the first connection the host and port of the last used server will be remembered for the next startup If during the next startup that server is not active anymore an error will be signaled and the user can connect to a different server instead 8 1 1 Starting a PowerLoom Server The GUI always connects to a PowerLoom server which can be embedded in the same Java process as in the examples above or can be run in a separate Lisp or Java process or as part of a Tomcat server a C Web server is not currently supported To start a PowerLoom server do the following Start the Java version of PowerLoom via the powerloom script or start a Lisp version of PowerLoom that supports a Web server such as Allegro CL or a Lisp that supports Portable AServe Once PowerLoom is up and running issue the following command PL U
123. ed to the file established by the last call to save all neural networks A number lt 0 or a NULL number turns off periodic saving sort self predicate FUNCTION CODE LIKE Method on QUERY SOLUTION TABLE SELF Perform a stable destructive sort of self according to predicate and return the result If predicate has a lt semantics the result will be in ascending order specializes subObject OBJECT superObject OBJECT Function TRUTH VALUE Try to prove if the description associated with subObject specializes the description for superObject and return the result truth value of the query start ontosaurus amp rest options OBJECT Command Start the PowerLoom HTTP server at port defaults to 9090 Loads the required support systems in Lisp and Java if necessary C is not yet supported Chapter 11 Miscellaneous 128 start powerloom gui amp rest options OBJECT Command Start the PowerLoom server at port defaults to 9090 and launches the GUI which will communicate with the server at that port If host is specified the GUI will try to communicate with a server at host port instead of the local embedded server note you can always point the GUI manually to a different server from its Connect to Server menu item Loads the required support systems if necessary Embedded calls to the GUI are currently only supported in Java however when the GUI is run in standalone mode it can communicate with any PowerLoo
124. em that company is a concept in the business module The second definition defines a concept corporation The type declaration c company indicates that corporation is a subconcept of company i e all instances of corporation are also instances of company Let us now create a couple of companies PowerLoom modules are case insensitive by default This means for example that a logical constant named Foo may be referenced by any of the symbols FOO foo foO etc You may create case sensitive modules but if you do so when inside that module all PowerLoom commands and other symbols such as AND EXISTS etc will need to be referred to using uppercase names since no automatic case conversion will take place Chapter 3 Annotated Example 12 assert company ACME cleaners assert corporation megasoft These two assertions create two new entities denoted by the terms ACME cleaners and megasoft Both of these entities are members of the concept company megasoft is also a member of the concept corporation We can test this by executing some PowerLoom queries retrieve all x company x There are 2 solutions 1 X ACME CLEANERS 2 X MEGASOFT retrieve all x corporation x gt There is 1 solution 1 X MEGASOFT 3 3 Relations So far our two companies aren t very interesting In order to say more about them we can define some relations and functions using the declarations defrelatio
125. ents happy and Fred Here is the list of all available PowerLoom commands add load path path srRING CONS OF STRING WRAPPER Command Append the directories listed in the separated path to the end of the STELLA load path Return the resulting load path 1 Lisp programmers are typically spoiled and find it inconvenient to wrap double quotes around their arguments Chapter 5 Commands 28 all facts of instanceRef OBJECT CONS OF PROPOSITION N Command Return a cons list of all definite TRUE or FALSE propositions that reference the instance instanceRef This includes propositions asserted to be true by default but it does not include propositions that are found to be TRUE only by running the query engine Facts inferred to be TRUE by the forward chainer will be included Hence the returned list of facts may be longer in a context where the forward chainer has been run then in one where it has not see run forward rules instanceRef can be a regular name such as fred as well as a function term such as father fred ask amp rest proposition amp options PARSE TREE TRUTH VALUE N Command Perform inference to determine whether the proposition specified in proposi tion amp options is true Return the truth value found ask will spend most of its effort to determine whether the proposition is true and only a little effort via shallow inference strategies to determine whether it is false To find out whether a proposi
126. ept populates the relation subpane with relations that use the concept as a domain type and populates the instance subpane with the concept s extension The proposition and rule subpanes are populated with propositions and rules associated with the concept e Selecting a relation populates the proposition and rule subpanes with propositions and rules associated with the relation e Selecting an instance with no selected relation populates the proposition subpane with propositions that refer to the selected instance e Selecting an instance and a relation populates the proposition subpane with proposi tions that contain the relation as a predicate and the instance as an argument e De selecting an object will update the state of the browser appropriately For example after selecting a module and a concept deselecting the concept will refresh the concept relation instance proposition and rule subpanes to display the knowledge contained in the selected module The title pane in each navigation pane displays a description of the source of the sub pane s contents For example if both the relation WINGSPAN and the instance AGM 130 were selected the proposition subpane would contain the title Propositions for WINGSPAN and AGM 130 Each selection event is recorded in a selection history which can be rolled back and forward For example assume the user selects the AIRCRAFT KB module and then selects the GUIDANCE TYPE concept If th
127. er of employees x n lt n 50 gt There is 1 solution 1 X ACME CLEANERS To repeat ourselves slightly Powerloom allows users the choice of using either relational or functional syntax when using a function in predicate position For example if f is a function then the expressions f x y and f x y are equivalent 3 6 Defined Concepts If we find ourselves writing the same query or subexpression repeatedly we may wish to define a name for the concept embodying that expression For example below we define the term small company to represent the class of all companies with fewer than 50 employees defconcept small company c company lt gt and Company c lt number of employees c 50 Notice that we have used a new keyword lt gt This keyword defines a bidirectional implication called if and only if Formally it is equivalent to the following pair of asser tions assert forall c gt and Company c lt number of employees c 50 small company c assert forall c gt small company c and Company c lt number of employees c 50 In other words the 7 keyword is a shortcut for an assertion that uses the lt gt relation which itself is a shortcut representing the conjunction of two single arrow implications For example lt gt P Q is equivalent to and lt P Q gt P Q where the lt relation is defined to be the inverse
128. erLoom GUI most closely resembles the traditional two tier client server model Since the GUI does not contain a great deal of business logic e g it does not know how to do inferencing it does not directly map onto the traditional notion of a smart client Similarly since PowerLoom is much smarter than a typical DBMS it does not cleanly map onto a traditional backend server However since the GUI contains the presentation logic it is more similar to a 2 tier model than a 3 tier model where the presentation logic resides in a middle tier Communication between the GUI and PowerLoom is done via the XML based SOAP pro tocol In order to effectively communication via SOAP a Web service layer was built on top of PowerLoom This layer provides support for marshaling and unmarshaling of PowerLoom objects to from XML and also provides a PowerLoom API that is accessible as a web ser vice The Java client uses JAXM and the Castor framework see http www castor org to support SOAP communication 8 1 Invoking the GUI The PowerLoom GUI can be started in a variety of ways The easiest way is to use the top level powerloom script see Section 1 3 Running PowerLoom page 4 and supply the gui option This will start the Java version of PowerLoom start a standalone embedded PowerLoom Web server launch the GUI application and connect from it to the PowerLoom server using port 9090 by default powerloom gui Running Java version of
129. erence to parent of One will sometimes see the command process definitions appearing at intervals within a file containing PowerLoom commands Each such appearance forces the definitions that preceed it to be fully evaluated This is done so that the interval between a definition and its evaluation not be too great it can get confusing if PowerLoom reports a semantic violation long after the origin of the conflict PowerLoom definitions commands those prefixed by def have one other semantic property that distinguishes them from ordinary assertions Any axioms that appear within a definition are tied to that definition If a definition is modified and then reevaluated axioms that don t survive the modification are retracted For example suppose we evaluate the following two commands defrelation parent of p1 person p2 person gt relative of p1 p2 1 The function and the concept number are predefined in PowerLoom Chapter 2 Conceptual Framework 8 defrelation parent of p1 person p2 person The first definition defines person as a binary relation and also states a rule that parent of implies relative of The second definitions erases that rule i e the cumu lative effect is as if the first definition did not appear In contrast consider the following commands defrelation parent of p1 person p2 person assert gt parent of p1 p2 relative of p1 p2 defrelation pa
130. eries iterative deepening Tells the query processor to use iterative deepening instead of a depth first search to find answers This is less efficient but necessary for some kinds of highly recursive queries trace subgoals Enables the generation of subgoaling trace information during backchaining inference trace solutions Prints newly found solutions during retrieval right when they are generated as opposed to when the query terminates trace classifier Tells the classifier to describe the inferences it draws justifications Enables the generation of justifications during inference which is a prerequiste for the generation of explanations with why emit thinking dots Tells PowerLoom to annotate its inference progress by out putting characters indicating the completion of individual reasoning steps By default the features emit thinking dots and just in time inference are en abled and the others are disabled set load path path STRING CONS OF STRING WRAPPER Command Set the STELLA load path to the separated directories listed in path Return the resulting load path time command command CONS OBJECT N Command Execute command measure and report its CPU and elapsed time needed for its execution and then return its result Chapter 5 Commands 39 unset feature amp rest features NAME LIST OF KEYWORD N Command Disable the PowerLoom environment feature s named by features Return the list of enab
131. f PowerLoom you can run PowerLoom from the top level PowerLoom directory using the powerloom script or powerloom bat under a Windows command prompt powerloom java Running Java version of PowerLoom Initializing STELLA Initializing PowerLoom Welcome to PowerLoom 4 0 0 Copyright C USC Information Sciences Institute 1997 2010 PowerLoom comes with ABSOLUTELY NO WARRANTY Type copyright for detailed copyright information Type help for a list of available commands Type demo for a list of example applications Type bye exit halt quit or stop to exit PL USER Similar to the C executable this will run various PowerLoom startup code and then bring up a PowerLoom command loop where you can execute commands Type demo to bring up a menu of available demos type run powerloom tests to run the PowerLoom test suite or type exit to exit PowerLoom 10 6 Removing Unneeded Files To save disk space you can remove files that you don t need For example if you are not interested in the C version of PowerLoom you can delete the directory native cpp Similarly you can remove native java to eliminate all Java related files You could do the same thing for the Lisp directory native lisp but in our opinion that would make it less Chapter 10 Installation 108 convenient for you to develop new PowerLoom code that is written
132. f actions that is relevant to the subpane that contains the mouse pointer The list of items will depend on whether the mouse is over a specific item or if it is over the background of the subpane s list or tree For example when the mouse is over a specific concept the menu will contain items for cutting pasting instantiating etc but when the mouse is over the background of the concept tree the only menu item presented will be to add a new concept The set of actions available for each subpane are as follows e Module Add Module Edit Module Load Module Save Module Clear Module Copy e Concept Add Concept Edit Concept Edit Extension Instantiate Cut Copy Paste Delete If multiple concepts are selected selecting Create New Concept from the background menu will create a concept that contains the selected concepts as parents e Relation Add Relation Edit Relation Edit Extension Copy Delete Show Inherited Direct Relations e Instance Add Instance Edit Instance Copy Delete Show Direct Derived Instances e Propositions Add Proposition Edit Proposition Copy Delete Navigate to Constituent Edit Constituent e Rules Add Rule Edit Rule Copy Delete Navigate to Constituent Edit Constituent 8 4 5 Editing Objects may be edited by right clicking them and selecting the Edit menu item in the popup menu Alternatively an object may be selected and then the Objects gt Edit Object menu item or the
133. ference level set to assertion Equivalent to query prop inference level assertion but more efficient BACKTRACKING QUERY prop PROPOSITION Relation Query prop with inference level set to backtracking Equivalent to query prop inference level backtracking but more efficient Chapter 7 Built In Relations 69 BINARY RELATION r RELATION Concept The class of binary relations BOUND VARIABLES arguments THING Relation True if all arguments are bound The bound variables predicate is used as a per formance enhancer to prevent other predicates from backchaining excessively while searching for bindings of certain of their arguments Purists will shun the use of this predicate but some rules are inherently inefficient without the addition of some kind of control logic Because evaluation of the bound variables predicate evaluation of predicates being guarded using this predicate has the side effect of locally dis abling query optimization See collect into set for an example that uses bound variables CARDINALITY c SET gt card INTEGER Function Function that returns the cardinality of a set CLOSED c COLLECTION Relation The collection c is closed if all of its members are known Asserting that a re lation is closed makes certain computations easier For example suppose that the relation happy is closed implying that all things that are happy will be asserted as such To prove not happy
134. fied name is required It is looked up using the Reflection tools The function found must conform to the signature for computation functions Arity specifies the number of arguments the computation accepts s register specialist function name STRING Function native name STRING module name STRING environment ENVIRONMENT Register name as a function name in the module named module name This function will the native code named native name The name is a fully qualified name which Chapter 11 Miscellaneous 133 will be interpreted by the normal rules for reading names in PowerLoom The native name will be processed in a manner that depends on the underlying programming language The following type mappings are used C Not available Error signaled Common Lisp The native name is read by READ FROM STRING and then the SYMBOL FUNCTION is taken Java A fully package qualified name is required It is looked up using the Reflection tools The function found must conform to the signature for specialist functions test environment level env ENVIRONMENT level STRING Function BOOLEAN Test if env has level set to level consify self CONS Method on PL ITERATOR Convert self into a Stella CONS listify self LIST Method on PL ITERATOR Convert self into a Stella LIST initialize kernel kb Command Bootstrap the PowerLoom built in kernel KB Chapter 12 Glossary 134 12 Glossary This glossary contains b
135. fies the relation relation retract proposition proposition PROPOSITION module MODULE Function environment ENVIRONMENT PROPOSITION Retract the truth of the proposition proposition in module Return the retracted proposition retract unary proposition relation LOGIC OBJECT arg OBJECT Function module MODULE environment ENVIRONMENT PROPOSITION Retract that the proposition relation arg is TRUE in module Return the asserted proposition Chapter 6 PowerLoom API 51 retrieve query CONS module MODULE environment ENVIRONMENT Function PL ITERATOR Returns an iterator of variable bindings that when substituted for the open variables in query satisfy the query proposition The query is run in module and relative to environment query has the same syntax as the PowerLoom retrieve command which see but with the retrieve operator omitted For example here are some legal query arguments happy x 10 happy x all happy x all x happy x 10 happy x inference level assertion 10 happy x inference level assertion timeout 1 0 If there is only a single output variable as in all the examples above each element generated by the returned iterator will be a binding for that variable unless the output variable was declared with a surrounding pair of parentheses For example all x happy x In that case the generated elements will be one element lists If there are multi
136. form actions relevant to the navigation pane Currently two toolbar buttons are present Add and Search Add indicated by a sign icon adds an object associated with the type of navigation pane Search indicated by a flashlight icon searches for objects associated with the type of navigation pane The content pane contains the actual knowledge to be displayed such as a list of instances or propositions There is one navigation pane for each type of KB object Modules Concepts Relations Instances Rules and Propositions Each internal pane is resizable by dragging the movable divider between the panes Panes may be hidden completely by clicking the Collapse arrow on the adjacent divider Clicking the Expand arrow will unhide the pane Visual Cues Navigation panes employ several visual cues to enhance the identifiability of object attributes Object types are indicated by an icon to the left of the object s name For example modules are represented by a blue M concepts by a red C etc The status of propositions is also indicated visually An italicized proposition indicates that the proposition was derived instead of asserted Grey propositions indicate that their truth value is a default value instead of a strict value Chapter 8 PowerLoom GUI 92 The main method for filling the contents of a navigation pane is to select some object in a navigation pane that is to the left or above it This is discussed in more detail in t
137. from the PLI ask function s ask does not expect a top level pair of parentheses For example here are some legal query arguments happy Fred happy Fred inference level assertion happy Fred inference level assertion timeout 1 0 Chapter 6 PowerLoom API 52 Names in query will be interpreted relative to module module name A null module name or the empty string refers to the current module If no module can be found with the name module name then a STELLA no such context exception is thrown The returned truth value represents the logical truth of the queried sentence as de termined by PowerLoom It can be be tested via the functions is true is false and is unknown which see s assert proposition sentence STRING module name STRING Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Assert that the logical sentence sentence is true in the module named module name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Return an iterator of the propositions resulting from sentence s change module name STRING environment ENVIRONMENT Function MODULE Set the current module to the module named name The return value is the module named name unless name is null or the empty string In that case the current module is returned If no module named name exists a Stella no such
138. function evaluate to each of them load stream stream INPUT STREAM environment ENVIRONMENT Function Read logic commands from the STELLA stream stream and evaluate them The stream should contain an in module declaration that specifies the module within which all remaining commands are to be evaluated The remaining commands are evaluated one by one applying the function evaluate to each of them load native stream stream NATIVE INPUT STREAM Function environment ENVIRONMENT Read logic commands from the native input stream stream and evaluate them As sumes stream is a line buffered stream which is a safe compromise but does not gener ate the best efficiency for block buffered streams such as files The stream should begin with an in module declaration that specifies the module within which all remaining commands are to be evaluated The remaining commands are evaluated one by one applying the function evaluate to each of them Chapter 6 PowerLoom API 50 next self BOOLEAN Method on PL ITERATOR Advance the PL Iterator self and return true if more elements are available false otherwise object to float self OBJECT FLOAT Function Coerce self to a float or throw a Stella Exception if the coersion is not feasible object to integer self OBJECT INTEGER Function Coerce self to an integer or throw a Stella Exception if the coersion is not feasible Floating point values will be coerced by rounding
139. gi da ace aoe ise 69 coerce to instance eee eee eee 114 coerce to instance or literal 114 COSTCE TO VeECtOr c acs weet acces MRRSSRS 114 COLLECT INTO ASCENDING SET 69 COLLECT INTO DESCENDING SET 69 COLEECT INT LIST 2 amp me RRERIHeY E REEEEPO 69 COLLECT INTO ORDERED SET 70 COLLECT INTO SED seis a4 kk IRR RES 70 collection 66 e aea eee RE BRPC ERAS 114 COLLECTIONOE 1e RR eave med 70 COMMENT zb E YRSLRUEOPERNEPUenS Uu EP ESSE 70 COMMUTATIVE 2 254 0 e280 4ceh ied TA Si el ee ie 70 CONCCIVE cedex ESEE EE EEEL ETETETT 30 41 Conceive term l2 pie derati ei eed 114 CONCEP Thoaan EE 7O CONCEPT PROTOTYPE s 2 24 30 seals oa cage hae emis TO conjoin truth values 00008 114 cons to pl iterator 22s esdeuiseceaneeses 41 consify on JUSTIFICATION 08 114 consify on PL ITERATOR 00005 133 consify on QUERY ITERATOR 0 114 consify on QUERY SOLUTION TABLE 114 consify current solutions on QUERY ITERATOR rp 114 consify justification 115 constamnmb 222 2 6 9 e E RR Ee Exe uus 115 copy on JUSTIFICATION 0 atri itani 115 COpyEiEght 29 10059 3bBPRDBEISRDO Mb ees 30 C VMEREN6 ei rette rr mer REOS RS ra Create oad uli eeu aen dutem Maen 115 create 2 d arrays esse e pees 115 create 2 d float array esses 115 ereate concept
140. gnaled The remaining commands are evaluated one by one applying the function evaluate to each of them load stream in module stream INPUT STREAM module MODULE Function environment ENVIRONMENT Read logic commands from the STELLA stream stream and evaluate them If the stream does not supply an in module declaration that specifies the module within which all remaining commands are to be evaluated it will be loaded in the module Chapter 11 Miscellaneous 132 specified If no module is specified and the file does not supply an in module decla ration an error will be signaled The remaining commands are evaluated one by one applying the function evaluate to each of them main Function Main PowerLoom entry point for your code in C and Java register computation function name STRING Function function reference FUNCTION CODE arity INTEGER module MODULE environment ENVIRONMENT Register name as a function name in module which will invoke the native code pro cedure described by function reference The name is a fully qualified name which will be interpreted by the normal rules for reading names in PowerLoom The function must conform to the signature for computation functions used by the computation specialist Arity specifies the number of arguments the computation accepts The exact form of function reference depends on the underlying programming lan guage The following type mappings are used C
141. he section below However in some cases it is possible to modify the contents of a navigation pane without performing a selection For example in the instance navigation pane it is possible to show derived or inherited instances by right clicking on the instance list and selecting an appropriate menu item Similarly the relation navigation pane can toggle between direct or inherited relations Propositions and rules are by default displayed according to the module that is currently selected However the contents of the proposition or rule navigation pane can be updated by selecting a more specific module in the View From Module combobox contained in the navigation pane s title bar Selection When the browser is initially opened a tree of modules is displayed in the module navigation pane and all other navigation panes are empty When a module is selected the remaining subpanes are populated with knowledge that is contained in that module Similarly selecting a concept in the concept navigation pane populates the relation proposition and instance panes with knowledge that is relevant to the selected concept In general selecting an object in a given navigation pane may affect the contents of navigation panes to the right and or below it More specifically the rules for object selection are as follows e Selecting a module populates the concept relation and instance subpanes with knowl edge contained in the module e Selecting a conc
142. he set of instances that belong to the concept Fact A fact is a proposition that has been asserted to be either True or False The term fact usually refers to a ground proposition i e a proposition that can be represented as a predicate applied to a sequence of instances or literals Filler The second argument to a binary tuple is often refered to as its filler When a multiple valued binary relation maps an instance to a set of values these values are also called fillers Forward Rule Function Formally a function is a relation such that the value of the last nth argument of a relational tuple is a function of the values of the first n 1 arguments This definition coincides with the notion of a single valued relation PowerLoom and KIF support specialized syntax that allows functions that have been defined using the operator deffunction to appear in term expressions e g f x 42 Instance An instance denotes an entity within a domain model a member of the concept Thing Depending on ones interpretation this could include almost everything Often the term instance is used more narrowly to exclude literals and other objects whose properties do not change over time PowerLoom assumes that concepts and relations are instances KIF Short for Knowledge Interchange Format KIF is a language that defines a Lisp like syntax for the predicate calculus There is an ANSII standard that
143. heir parents specified via the includes option and or a fully qualified module name and 2 they inherit from used modules specified via the uses option The main difference between the two mechanisms is that inheritance from parents is Chapter 5 Commands 32 transitive while uses links are only followed one level deep I e a module A that uses B will see all objects of B and any of B s parents but not see anything from modules used by B Another difference is that only objects declared as public can be inherited via uses links this is not yet enforced Note that contrary to Lisp there are separate name spaces for classes functions and variables For example a module could inherit the class CONS from the STELLA module but shadow the function of the same name The above discussion of includes and uses semantics keyed on the inheri tance visibility of symbols The PowerLoom system makes another very important distinction If a module A is inherited directly or indirectly via includes specification s by a submodule B then all definitions and facts asserted in A are visible in B This is not the cases for uses the uses options does not impact inheritance of propositions at all The list of modules specified in the includes option plus if supplied the parent in the path used for name become the new module s parents If no uses option was supplied the new module will use the STELLA module by default otherwise it wi
144. hod on QUERY SOLUTION TABLE Remove and return the first solution of self or NULL if the table is empty post for evaluation self PRoPOSITION world CONTEXT Function Push self onto the evaluation queue unless it s already there powerloom Function Run the PowerLoom listener Read logic commands from the standard input evaluate them and print their results Exit if the user entered bye exit halt quit or stop powerloom gui exit hook ignore OBJECT Function Exit hook to stop the PowerLoom GUI if it is running powerloom information STRING Command Returns information about the current PowerLoom implementation Useful when reporting problems pretty print logical form form OBJECT Function stream OUTPUT STREAM Pretty print the logical form form to stream according to the current setting of logic dialect Chapter 11 Miscellaneous 124 print array self stream NATIVE OUTPUT STREAM Method on 2_D_ARRAY Print the array self to stream print array self Method on 2_D_FLOAT ARRAY stream NATIVE OUTPUT STREAM Print the array self to stream print extension sizes module MODULE sizeCutoff INTEGER Function Print the extension sizes of concepts visible in module If module is NULL the current module is used Do not report extensions with size less than sizeCutoff default is 10 print facts instanceref OBJECT N Command Like ALL FACTS OF but prints each fact on a sep
145. i stella Iterator object in their constructor Chapter 7 Built In Relations 68 7 Built In Relations This chapter lists relations that come predefined in PowerLoom They are defined in the module PL KERNEL users can access them by including or using the PL KERNEL module within the declarations of their own modules x NUMBER y NUMBER gt z NUMBER Function Function that multiplies two numbers x NUMBER y NUMBER gt z NUMBER Function Function that adds two numbers x NUMBER y NUMBER gt z NUMBER Function Function that subtracts two numbers x NUMBER y NUMBER gt z NUMBER Function Function that divides two numbers lt x THING y THING Relation True if x lt y lt x THING y THING Relation True if x lt y gt x THING y THING Relation True if x gt y gt x THING y THING Relation True if x gt y ABSTRACT r RELATION Relation rue if there are no direct assertions made to the relation r AGGREGATE a AGGREGATE Concept 7a is an aggregate ANTISYMMETRIC r RELATION Relation A binary relation r is antisymmetric if whenever r x y is true r y x is false unless x equals y ARITY r RELATION gt arity INTEGER Function The number of arguments domains of the relation r ASSERTION QUERY prop PROPOSITION Relation Query prop with in
146. ic Interface This section contains the description of the programming language specific aspects of using the PowerLoom API Each section describes the naming conventions and namespace issues related to calling the API functions from that programming language Chapter 6 PowerLoom API 59 6 2 1 Lisp API This section tells how to call the API functions in PowerLoom s Common Lisp implemen tation from a Lisp program The function names are identical to the Stella names in the PowerLoom API description See Chapter 6 PowerLoom API page 40 They are exported from the PLI package Other Stella symbols and names are in the STELLA package but currently none of the Stella symbols are exported PowerLoom can be used from Allegro Common Lisp CMU Common Lisp LispWorks Common Lisp and Macintosh Common Lisp It may be possible to use the system from other Common Lisp systems but they have not been tested 6 2 1 1 Common Lisp Initialization Loading the Common Lisp version of PowerLoom will normally initialize the system as part of the loading process The Common Lisp version can be loaded by loading the file Load powerloom lisp from the top level powerloom directory This will make the system available for use 6 2 1 2 Type Declarations Stella is a typed language and the Common Lisp translation uses the type information for Common Lisp type declarations That means that values specified as being of type INTE GER STRING and FLOAT
147. ing Modules We begin by creating a PowerLoom module which is a logical container that holds the term definitions rules facts etc that make up all or a portion of a domain model We will call our module business The defmodule command defines a new module The includes option within the defmodule tells PowerLoom that the business module in herits all definitions and assertions present in the PL USER module or in ancestor modules inherited by the PL USER module In particular by inheriting PL USER we indirectly in herit the PL KERNEL module that contains all of the built in concepts and relations The in module command tells the PowerLoom system to make BUSINESS the current module Until the current module is changed again all new introductions of terms and facts will be placed in the business module defmodule BUSINESS includes PL USER in module BUSINESS The basic building blocks of a model are its concepts relations and instances A concept defines classes categories of entities that populate the domain model A relation defines attributes and relationships that allow the declaration of facts about an entity Instances are members of concepts They appear as arguments to propositional assertions 3 2 Concepts Concepts are defined using the defconcept command Here we define the concepts company and corporation defconcept company defconcept corporation c company The first definition tells the syst
148. ionally multiple knowledge browsers can be open at the same time to present different views of a KB The Knowledge Browser itself consists of multiple collapsible and resizable subpanes which in turn are composed of scrollable lists and trees This allows a birds eye view of a Powerloom KB in which many modules concepts relations instances propositions and rules can be displayed at the same time Navigability When exploring a KB it is imperative that a user interface allows easy navigation between related objects The PowerLoom GUI has extensive navigation capabil ities which are as good or better than browser based applications For example a user may click on a query result to instantly update the Knowledge Browser to display the selected object Also a user can right click on a relation or argument in a proposition and navigate to the clicked on object Responsiveness For the best possible user experience a user interface should be highly responsive to a user s input gestures For example 1 after initiating a gesture such as a mouse click or keypress there should be a minimal delay before the appliction performs the intended action and 2 Power Users should be able to perform complex tasks with a minimum number of mouse clicks key presses etc The PowerLoom GUI attempts to minimize network round trips by caching large amounts of data For example when a user points the Knowledge Browser to a module a large chu
149. is used A zero column number returns the proposition s relational predicate If the return value is not a LOGIC OBJECT an exception is thrown As a special case a column number of zero will also return sequence itself if it is not one of the types enumerated above This is done to allow the use of get nth value on PL iterators with only a single return variable If sequence is not a LOGIC OBJECT an exception is thrown get nth string sequence OBJECT n INTEGER module MODULE Function environment ENVIRONMENT STRING Return a string representation of the value in the nth column of sequence Counting starts at zero Unless n is zero sequence must be of type proposition cons vector or PL iterator A zero column number returns a proposition s relational predicate For the PL iterator case the the current value pointed to by the iterator is used This will always succeed even if the nth value is not a string object In that case a string reprensentation will be returned As a special case a column number of zero will also return sequence itself as a string if it is not one of the types enumerated above This is done to allow the use of get nth string on PL iterators with only a single return variable get nth value sequence OBJECT n INTEGER module MODULE Function environment ENVIRONMENT OBJECT Return the value in the nth column of sequence Counting starts at zero Unless n is zero sequence must be of type propos
150. ist undefined relations which see call propagate constraints context CONTEXT Function Trigger constraint propagation over all propositions in the module or world context call retrieve query OBJECT QUERY ITERATOR Function Callable version of retrieve which see Accepts queries specified by a query itera tor or specified as a CONS list of arguments as they would be supplied to retrieve Raises LOGIC EXCEPTIONSs in case of illegal queries and logical expressions call retrieve partial query OBJECT QUERY ITERATOR Function Callable version of retrieve partial which see Accepts queries specified by a query iterator or specified as a CONS list of arguments as they would be supplied to retrieve partial Raises LOGIC EXCEPTIONS in case of illegal queries and logical expressions call run forward rules module MODULE force BOOLEAN Function Run forward inference rules in module module If module is NULL the current module will be used If forward inferencing is already up to date in the designated module no additional inferencing will occur unless force is set to T RUE in which case all forward rules are run or rerun Chapter 11 Miscellaneous 114 call set inference level levelKeyword KEYWORD Function module MODULE KEYWORD Set the inference level of module to the level specified by levelKeyword If module is NULL and we are inside a query set the level of the current query iterator Otherwise
151. ith intractable and expensive reasoning procedures PowerLoom is very scalable and comes with a variety of mechanisms to control search and handle large ontologies and knowledge bases For example PowerLoom s reasoning specialists han dle a wide variety of inferencing very effectively without having to go through any rule chaining Search control annotations can be used to help the inference engine use rules more effectively For example depending on fan out certain rules are run more effectively forwards than backwards or vice versa and a KB developer can tell the system when that s the case PowerLoom has resource bounded depth first or iterative deepening search which provides an any time inference scheme for situations where resources are limited A just in time forward inference engine elaborates the asser tion neighborhood of objects touched upon by inference This allows focused forward inference relevant to current inference goals without having to run forward inference over a potentially very large KB to completion PowerLoom has a static and dynamic query optimizer that similar to optimizers used in database systems orders conjunc tive goals based on relation extension sizes and rule fan out to minimize intermediate result sets and chaining The dynamic optimizer does this for each conjunctive sub goal based on actual bindings Given this mechanism it is possible to run PowerLoom queries that return 100 000 s of solutions Powe
152. ition cons vector or PL iterator A zero column number returns a proposition s relational predicate For the PL iterator case the number of columns is for the current value of the iterator As a special case a column number of zero will also return sequence itself if it is not one of the types enumerated above This is done to allow the use of get nth value on PL iterators with only a single return variable get operator name STRING SYMBOL Function Returns the logical operator object a Stella SYMBOL for name If no such operator exists then a no such object exception is thrown get object name STRING module MODULE Function environment ENVIRONMENT OBJECT Look for an object named name that is local to or visible from the module module get parent modules module MODULE PL ITERATOR OF MODULE Function Return the modules that are immediate parents of module get predicate prop PROPOSITION LOGIC OBJECT Function Return the concept or relation predicate for the proposition prop Chapter 6 PowerLoom API 47 get proper subrelations relation LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF LOGIC OBJECT Return relations that specialize relation Non reflexive get proper superrelations relation LOGIC OBJECT Function module MODULE environment ENVIRONMENT PL ITERATOR OF LOGIC OBJECT Return relations that generalize relation Non reflexive get propositi
153. iz8 2 so ag eat ease ede pr ES 48 initialize kernel kb 133 insert at on QUERY SOLUTION TABLE t21 INSERT ELEMENT oe ic ede tess Rn RR aseri 74 INSTANCEOF oase e pg der eed er oea ERDAN 74 INVERSE de RISE RE RES PARRRPRE MEE RE 74 Function Index invert truth value esses ees 121 IRREELEXIVE 2 22 9 Gk bee Res b e TA e E T EE id hi diets ahd Guin EEE 48 DSH CSF AWG nescrise renano neu dea RO dob teense 48 is enumerated collection 48 is enumerated list 0 cece eee ee 48 is enumerated set esee 48 18 5falS6 2 bed re Pb RR arate 48 J8 flO8t 22 ease RU HER hn RR Ed 48 is inconsistent 2 u khi ss nee Ri EGER 48 is integer xnrgerdoveeeatema Hurra 48 DS KNOW 44 ees tee sordes dre bala Sacs uh RAP E Pre 131 is logic objeCt 20 gites enere 48 DSANUMDEE i e decane thes Gd se uad dant hagas REP a 48 JS6 SUElCb v gases aH bap rebat E RE Geb don ae 48 d6 SLEIBE Lnesessni ida ck eia Ru tee P OLER RR Ran 48 iS subrelatiOn il lo nates e aed medela 49 18 tfr 6 lilbudoesnd adgressi EET 49 is true binary proposition 49 is tru e proposition e rz een ug 49 is tr e propositionl e meet 131 is true unary proposition 49 286 UnknOWLD 22r E eg d re dob ker 49 ISSUE eed me Rieder e ns vs RES esee en TRUE 74 EISE secnndo arate dang eee ane gue Ea enone 74 iterator to pl iterator 5 c ee i 49
154. kisweieeq CLE See eL ERU TT pop on QUERY SOLUTION TABLE sss ees 123 pop rioad path i eire ERR I 35 post for evaluation i xciuuuwi eine pe 123 POWCPLOOM voi a cceccecte ican whe ad tetas aside 123 powerloom gui exit hook 123 powerloom information sss 123 Function Index presume risiede eds 35 pretty print logical form 123 print array on 2_D_ARRAY 005 124 print array on 2 D FLOAT ARRAY 124 print extension sizes s sss 124 PLING LACS sie 1 lh Dg UU RP IOS 124 print feat res a Lg erkm s bee hie iad 35 print goal stack v volvo epar es 124 print logical form 2x e e re nexa 124 print logical form in dialect 124 print ruleS 542 m Dm reb n eb 35 50 print unformatted logical form 124 print whynot justification 124 process definitiohns i L6 iylel mes 35 PROJECT CUEUMN encre Gadees cee ER Tif propaga te constreint8 s c Redes 36 PROPER SUBRELATIUN nets Af PROPER SUPERRELATEION i m Ren aa PROPOSITION ARGUMENT L eR RIDES 78 PROPOSITION ARGUMENTS 006 ec rs 78 PROPOSITION ARITY 236 os 9 mee 78 PROPOSITION RELATION occ 3 einen coeds ace 78 push load path sirrorrsiciyrrreni rennan ma 36 Q QUERY ien cs eesce temere EG eeR N audio 78 random fl 8 t 2 9 e pee Gade ve ped 124 RANGE ecd spa eR bores Cicadas der eeeaie sae 78 RANGE CARDINALITY
155. l need to be recompiled manually better patch recompilation support might become available in the future Chapter 11 Miscellaneous 109 11 Miscellaneous This is a catch all section for documented functions methods and relations that haven t been categorized yet into any of the previous sections They are in random order and many of them will never be part of the official PowerLoom interface So beware 2_d_element array row INTEGER column INTEGER Method on 2_D_ARRAY LIKE ANY VALUE SELF Return the element of array at position row column 2_d_element array row INTEGER Method on 2_D_FLOAT ARRAY column INTEGER FLOAT Return the element of array at position row column 2_d_element setter array value OBJECT Method on 2_D_ARRAY row INTEGER column INTEGER LIKE ANY VALUE SELF Set the element of array at position row column to value and return the result 2_d_element setter array value FLOAT Method on 2_D_FLOAT ARRAY row INTEGER column INTEGER LIKE ANY VALUE SELF Set the element of array at position row column to value and return the result add testing example form CONS score PARTIAL MATCH SCORE N Command Add a query and score pair to the master list of testing examples add training example form CONS N Command score PARTIAL MATCH SCORE Add a query and score pair to the master list of training examples all asserted types self OBJECT CONS OF NA
156. ld values extraneous values will be ignored fill array self amp rest values FLOAT Method on 2_D_FLOAT ARRAY Fill the two dimensional array self in row major order from values Missing values will retain their old values extraneous values will be ignored finalize objects Function Finalize all currently unfinalized objects The user level entry point for this is process definitions Chapter 11 Miscellaneous 119 find direct supers and subs self DESCRIPTION Function onlysupers BOOLEAN CONS OF DESCRIPTION CONS OF DESCRIPTION CONS OF DESCRIPTION Classify self and return three values its direct supers direct subs and a list of equiv alent descriptions Setting supersOnly may speed up the computation perhaps by a lot If description is nameless and has no dependent propositions then it is automatically removed from the hierarchy after classification find direct supers of instance self OBJECT CONS OF Function LOGIC OBJECT Classify self and return a list of most specific named descriptions among all descrip tions that it satisfies find instance instanceRef OBJECT OBJECT N Command Return the nearest instance with name instanceRef visible from the current module instanceRef can be a string symbol or surrogate If instanceRef is a surrogate the search originates in the module the surrogate was interned in find rule ruleName NAME PROPOSITION N Command Search for a rule nam
157. le name then a Stella no such context exception is thrown s print rules name STRING stream OUTPUT STREAM Function module name STRING environment ENVIRONMENT Print rules attached to the concept relation named name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s retract proposition sentence STRING module name STRING Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Retract the truth of the logical sentence sentence in the module named module name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Return an iterator of the retracted propositions resulting from sentence s retrieve query STRING module name STRING Function environment ENVIRONMENT PL ITERATOR Returns an iterator of variable bindings that when substituted for the open variables in query satisfy the query proposition The query is run in module and relative to environment query has the same syntax as the PowerLoom retrieve command which see but with the retrieve operator omitted Different from the PLI retrieve function s retrieve does not expect a top level pair of parentheses For example here are some legal query arguments happy x 10 happy x all happy
158. led features Calling unset feature without any arguments can be used to display the currently enabled features See set feature for a description of supported features why amp rest args OBJECT N Command Print an explanation for the result of the most recent query Without any arguments why prints an explanation of the top level query proposition down to a maximum depth of 3 why all prints an explanation to unlimited depth Alternatively a particular depth can be specified for example why 5 explains down to a depth of 5 A proof step that was not explained explicitly e g due to a depth cutoff can be explained by supplying the label of the step as the first argument to why for example why 1 2 3 5 prints an explanation starting at 1 2 3 down to a depth of 5 which is counted relative to the depth of the starting point The keywords brief and verbose can be used to select a particular explanation style In brief mode explicitly asserted propositions are not further explained and indicated with a assertion marker Additionally relatively uninteresting proof steps such as AND introductions are skipped This explanation style option is sticky and will affect future calls to why until it gets changed again The various options can be combined in any way for example why 1 2 3 brief 3 explains starting from step 1 2 3 down to a depth of 3 in brief explanation mode Chapter 6 PowerLoom API 40 6 PowerLoom API This chapte
159. ling FALSE as the optional second argument or by typing c at the pause prompt Typing at the pause prompt prints a list of available commands deny proposition PARSE TREE OBJECT N Command Assert the falsity of proposition Return the asserted proposition object KIF exam ple deny happy Fred asserts that Fred is not happy which could have been done equivalently by assert not happy Fred Note that for this to succeed the relation happy must already be defined see assert describe name OBJECT amp rest mode OBJECT N Command Print a description of an object in verbose terse or source modes destroy objectSpec PARSE TREE OBJECT N Command Find an object or proposition as specified by objectSpec and destroy all propositions and indices that reference it objectSpec must be a name or a parse tree that evaluates to a proposition Return the deleted object or NULL if no matching object was found drop load path path STRING CONS OF STRING WRAPPER Command Remove the directories listed in the separated path from the PowerLoom load path get load path CONS OF STRING WRAPPER Command Return the current STELLA load path get rules relation NAME CONS OF PROPOSITION N Command Return the list of rules associated with relation help amp rest commands SYMBOL N Command Describe specific commands or print a list of available commands Chapter 5 Commands 35 in mod
160. ll use the set of specified modules If case sensitive is supplied as TRUE symbols in the module will be interned case sensitively otherwise the default they will be converted to uppercase before they get interned That means that any reference from inside a case sensitive module to a non case sensitive module will have to use uppercase names for symbols in the non case sensitive module The standard system modules are all NOT case sensitive Modules can shadow definitions of functions and classes inherited from parents or used modules Shadowing is done automatically but generates a warning unless the shadowed type or function name is listed in the shadow option of the module definition Examples defmodule PL KERNEL PL USER uses LOGIC STELLA package PL USER defmodule PL USER GENEALOGY The remaining options are relevant only for modules that contain STELLA code Modules used only to contain knowledge base definitions and assertions have no use for them The keywords lisp package java package and cpp package specify the name of a native package or name space in which symbols of the module should be allocated when they get translated into one of Lisp Java or C By default Lisp symbols are allocated in the STELLA package and C names are translated without any prefixes The rules that the STELLA translator uses to attach translated Java objects to classes and packages are somewhat complex Use java pack
161. lla and by extension PowerLoom is organized a little bit differently than in Java You can either use the Stella iterators directly or else use one of the wrapper classes described in the section Section 6 2 3 10 Utility Classes for Java page 67 It will present a more familiar Java interface Since the iteration models are a bit different it would be unwise to mix accesses between the iteration models Stella iterators do not compute any values until the next method in Java nextP is called This method will try to compute the next value of the iterator and it will return a boolean value which is true if more values are present Each time it is called the iteration advances Values can be read out of the value field of the iterator which will have type Stella Object Some iterators will also have a key field which can be read The way one would normally use a Stella iterator is as follows with possible casting of the value field Stella Iterator iter while iter nextP processValue iter value j The PLI class also contains a number of functions for simplifying the interaction with PlIterator objects that are returned by various API functions These are the getNth functions They work on sequences and sequence like objects such as Proposition and and the value slot of PlIterator objects Note that they do not return sequential elements of the iterator but rather elements of the implicit sequence that is the value
162. m installation that supports an HTTP server currently Lisp and Java start powerloom server amp rest options OBJECT Command Start the PowerLoom HTTP server at port defaults to 9090 Loads the required support systems in Lisp and Java if necessary C is not yet supported stop ontosaurus Command Stop the PowerLoom HTTP server and free up any bound ports This is a no op if no server is running or the server is not supported stop powerloom gui Command Closes the PowerLoom GUI application if it is currently visible This is a no op if the GUI is not running or if it is not supported stop powerloom server Command Stop the PowerLoom HTTP server and free up any bound ports This is a no op if no server is running or the server is not supported strengthen truth value tvl TRUTH VALUE tv2 TRUTH VALUE Function TRUTH VALUE If tv2 has greater strength than tv1 adapt the strength of tv1 not its value and return the result Otherwise return tv1 unmodified strict truth value self TRUTH VALUE BOOLEAN Function Return TRUE if self is a strict truth value termify self OBJECT OBJECT Function Convert self into an equivalent PowerLoom object that can be passed as an argument wherever an instance is expected test closed slot relation SURROGATE BOOLEAN Function Return TRUE if relation a surrogate is asserted to be closed or if the current module closes all relations tes
163. members of A Truth Maintenance Type Often used a synonym for the term concept The phrase a type of an instance generally refers to one of the concepts that the instance belongs to The phrase nth domain type refers to a concept that contains all instances of the nth column of a relation World Chapter 13 PowerLoom Grammar 137 13 PowerLoom Grammar The syntax of PowerLoom is described below using a modified BNF notation adapted from the KIF specification 13 1 Alphabet We distinguish between terminals which are part of the language and nonterminals All nonterminals are bracketed as follows lt nonterminal gt Squared brackets means zero or one instances of the enclosed expression lt nonterminal gt means zero or more occurrences and lt nonterminal gt means one or more occurrences of lt nonterminal gt The notation lt nonterminali gt lt nonterminal2 gt refers to all of the members of lt nonterminal1 gt except for those in lt nonterminal2 gt A word is a contiguous sequence of characters which include all upper case letters lower case letters digits and alpha characters ASCII character set from 93 to 128 excluding some special characters like white spaces single and double quotes and brackets lt word gt a primitive syntactic object Special words are those who refer to a variable All variables are preceded by a question mark lt indvar gt a word beginning with the charac
164. ment ENVIRONMENT LOGIC OBJECT Create a function named name with arity arity in the designated module Domain and range information can be added via assertions of nth domain or domain and range relations Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly create module name STRING parent MODULE Function case sensitive BOOLEAN MODULE Creates a new module name as a child of parent The flag case sensitive controls whether names read in this module will be case sensitive or not create object name STRING concept LOGIC OBJECT Function module MODULE environment ENVIRONMENT LOGIC OBJECT Create an object named name of type concept in the designated module Both name and concept can be null If name is null then an object will be created with a new non conflicting name based on the name of concept or system generated if no concept is specified If concept is nu11 then the object will be of type THING It is an error to create an object with the same name as an existing object Note that the string can be a qualified name in which case the object will be created in the module specified but with a name as determined by the qualified name Vertical bars in the name string are interpreted as Stella escape characters Note that because names in modules that are not case sensitive are canonicalized the name of the returned object m
165. mentioned below will be relative to that directory which we will usually refer to as the PowerLoom directory 10 3 Lisp Installation To install the Lisp version of PowerLoom startup Lisp and load the file load powerloom lisp with CL load load powerloom lisp The first time around this will compile all Lisp translated STELLA files before they are loaded During subsequent sessions the compiled files will be loaded right away By default PowerLoom now uses the version that uses Lisp structs instead of CLOS objects to implement STELLA objects If you do want to use the CLOS based version instead do the following CL setq cl user load cl struct stella CL nil CL 1oad load powerloom lisp Alternatively you can edit the initial value of the variable load cl struct stella in the file 1o0ad powerloom lisp Using structs instead of CLOS objects greatly improves slot access speed however it may cause problems with incremental re definition of STELLA classes this is only an issue if you are developing your application code in the STELLA language In that case it is recommended to only use the struct option for systems that are in or near the production stage Once all the files are loaded you should see a message similar to this PowerLoom 3 2 0 loaded Type powerloom to get started Type in package STELLA to run PowerLoom commands directly from the Lisp top level USER 2 To reduce startu
166. menu items for showing hiding the application s toolbar and status bar This menu also contains items for changing the application s font for example the Demo Theme changes all fonts to a large bold font suitable for demo presentations e Navigate contains menu items for navigating backward and forward in a browser s selection history e Window contains a list of the currently open windows on the desktop Selecting a window brings the window to the top of the window stack and if the window is iconified it is de iconified e Help contains an item to open an HTML help browser and an item to open an About box which contains information about the PowerLoom GUI Most menu items have accelerator keys that allow an item to be executed by a combina tion of keystrokes The toolbar contains several buttons which provide shortcuts to menu items There are currently toolbar buttons for cutting copying pasting deleting editing an object and opening a query dialog The toolbar may be undocked from its default position by dragging it anywhere on the desktop It may also be hidden and unhidden by selecting the View gt Toolbar menu item The status bar at the bottom of the application contains information on the current status of the application The status bar is divided into two sections The leftmost section displays the last module that was selected by a user The application keeps track of the current module in order to provide conti
167. mode without running an interactive command loop This can be useful in conjunction with the eval option to execute a single command or load a PowerLoom script via the load command help Prints a list of all currently supported command line options Chapter 2 Conceptual Framework 6 2 Conceptual Framework This chapter presents the fundamental conceptual building blocks that are used to construct PowerLoom knowledge bases The PowerLoom language is based on KIF which provides a syntax and a declarative semantics for first order predicate calculus expressions KIF is a proposed ANSII standard language used by a variety of knowledge representation systems Practical knowledge representation systems necessarily add a procedural semantics that defines the interpretation of knowledge structures when definitions and facts are retracted or modified This chapter assumes that the reader has some familiarity with the semantics of the predicate calculus and instead focuses on aspects of the semantics that go beyond the traditional KIF semantics A PowerLoom knowledge base is constructed by first defining the terminology concepts and relations for a domain and then asserting additional rules and facts about that domain Facts can be asserted and later retracted so the answers returned by queries may change over time The knowledge structures are organized into logical containers called modules The division into modules means that in gener
168. moduleRef Chapter 11 Miscellaneous 120 get relation instanceRef OBJECT LOGIC OBJECT Function Return the nearest relation with name instanceRef visible from the current module instanceRef can be a string symbol or surrogate If instanceRef is a surrogate the search originates in the module the surrogate was interned in get self or prototype instanceRef OBJECT LOGIC OBJECT Function Used to convert a computation to reference so called template slots rather than own slots If instanceRef denotes a class return a prototype of that class Otherwise return instanceRef get slot maximum cardinality self LOGIC OBJECT Function relation SURROGATE INTEGER Return a maximum value for the number of fillers of relation relation a surrogate applied to the instance self an object get slot minimum cardinality self LOGIC OBJECT Function relation SURROGATE INTEGER Return a minimum value for the number of fillers of relation relation a surrogate applied to the instance self an object get slot value self LOGIC OBJECT relation SURROGATE OBJECT Function Return a single value for the slot relation a surrogate applied to self an object get slot value type self LoGIC OBJECT relation SURROGATE Function NAMED DESCRIPTION Return a most specific type for fillers of the slot relation a surrogate applied to self If there is more than one pick one get why justification label STRING
169. multiple language to allow easy and effective integration into knowledge based applications PowerLoom s focus is on expressivity of its representation language while still providing good scalability to large ontologies and knowledge bases In general PowerLoom takes a pragmatic stance where usability is more important than theoretical neatness and ex pressivity is more important than inferential completeness From our point of view there is nothing magical about logic it is just another programming language with difficult to understand semantics so it should help you to solve the task at hand as best as possible and not hinder you by forcing you to work around restrictions of the logic Of course PowerLoom cannot completely escape the straight jacket of logic but it tries to push the boundaries as much as possible e Representation language PowerLoom uses the language of predicate logic to represent knowledge The syntax is KIF the Knowledge Interchange Format which is one of the supported syntaxes of the upcoming Common Logic standard PowerLoom adds a variety of convenient definitional constructs as well as extensions beyond traditional first order logic such as type level predicates relation variables in holds sentences modal assertions sentences about sentences cross context assertions via ist to rep resent lifting axioms defaults still experimental and others The goal is to provide a highly expressive representation language
170. n s get inferred binary proposition values Function relation name STRING arg name STRING module name STRING environment ENVIRONMENT PL ITERATOR Return all values v such that relation name arg name v has been asserted or can be inferred A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get nth domain relation name STRING n INTEGER Function module name STRING environment ENVIRONMENT LOGIC OBJECT Return the type a concept for the nth argument of the relation named relation name Counting starts at zero A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get object name STRING module name STRING Function environment ENVIRONMENT OBJECT Look for an object named name that is local to or visible from the module module name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get parent modules name STRING environment ENVIRONMENT Function PL ITERATOR OF MODULE Return the modules that are immediate parents of module name If no module named name exists a Stella no such context exception is thrown s get parent m
171. n and deffunction defrelation company name c company name STRING This declaration defines a binary relation company name The first value in a company name tuple must be an instance of type company while the second value must be a string We can now give our companies names using the command assert assert company name ACME cleaners ACME Cleaners LTD assert company name megasoft MegaSoft Inc We can retrieve pairs of companies and their names with the following query note that we omitted the optional retrieval variables in which case they are determined by collecting the free variables in the query expression retrieve all company name x y There are 2 solutions 1 X MEGASOFT Y MegaSoft Inc 2 X ACME CLEANERS Y ACME Cleaners LTD Using retrieval variables is useful if we want to order the result columns in a certain way for example retrieve all y x company name x 7y There are 2 solutions 1 Y MegaSoft Inc X MEGASOFT 2 Y ACME Cleaners LTD X ACME CLEANERS Chapter 3 Annotated Example 13 3 4 Relation Hierarchies PowerLoom permits the specification of hierarchies both for concepts and relations Previ ously we defined a small concept hierarchy with company on top and corporation below it We now define a subrelation of the relation company name called fictitious business name defrelation fictitious business name c company name STRING
172. n deleted will be considered If local only return propositions conceived locally in module all relation values relation SURROGATE Function nMinusOneArguments CONS CONS Return a set of values that satisfy the relation relation a surrogate applied to nMi nusOneArguments plus that last value all sentences of instanceRef OBJECT CONS OF Command STRING WRAPPER Return a list of sentences describing facts about instanceHef all slot value types self LOGIC OBJECT relation SURROGATE Function CONS OF NAMED DESCRIPTION Return a set of the most specific types for fillers of the slot relation applied to self all slot values self LOGIC OBJECT relation SURROGATE CONS Function Return a set of values for the slot relation a surrogate applied to self an object all subrelations relation NAMED DESCRIPTION Function removeequivalents BOOLEAN CONS OF NAMED DESCRIPTION Return a set of all named relations that specialize relation all superrelations relation NAMED DESCRIPTION Function removeequivalents BOOLEAN CONS OF NAMED DESCRIPTION Return a set of all relations that subsume relation all taxonomic types self OBJECT CONS OF NAMED DESCRIPTION Function Return a set of all of the types that are satisfied by self using only assertions and upward taxonomic reasoning all terms module MODULE local BOOLEAN ITERATOR OF Function OBJECT Return a list of all term
173. n depends on the value of ifexists Possible values are ASK REPLACE WARN and ERROR REPLACE gt Means overwrite without warning WARN gt Means overwrite with a warning ERROR gt Means don t overwrite signal an error instead ASK gt Ask the user whether to overwrite or not If not overwritten an exception is thrown A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown save module module MODULE filename STRING ifexists STRING Function environment ENVIRONMENT Save the contents of the module mod into a file named filename If a file named filename already exists then the action taken depends on the value of ifexists Possible values are ASK REPLACE WARN and ERROR REPLACE gt Means overwrite without warning WARN gt Means overwrite with a warning ERROR gt Means don t overwrite signal an error instead ASK gt Ask the user whether to overwrite or not If not overwritten an exception is thrown string to object string STRING type LOGIC OBJECT Function module MODULE environment ENVIRONMENT OBJECT Evaluate string with respect to module and environment and return the corresponding logical term type is a concept used to assist the correct interpretation of string Currently type only has an effect on the interpretation of literal types 6 2 Language Specif
174. nd I In that case R is viewed as the canonical relation and I simply provides a different access mechanism to the canonical relation In a logic based KR paradigm inverse relations are redundant and do not add anything that couldn t be represented or queried without them however sometimes they can provide some extra convenience for users Asserting inverse I R also will not cause an error but can degrade backward inference performance due to the extra redundant rule that gets generated If domain rules will be written in terms of both R and I as opposed to only R inverse I R should be asserted also to get full inferential connectivity between the two relations IRREFLEXIVE r RELATION Relation A binary relation r is irreflexive if it is false when both of its arguments are identical ISSUE x THING 7s STRING Relation 7s is an issue attached to x An issue string normally comments on a topic that has not been resolved to everyone s satisfaction IST context CONTEXT p PROPOSITION Relation True if proposition p is true in context context The IST is true relation allows one to evaluate a query or rule in more than one context A common use of IST is in defining lifting axioms that import knowledge from one context to another For example below is a rule that accesses a patient record relation in a module called Medical Kb lifts out the age column and imports it into a has age relation in the current context
175. nd relation hi erarchies and instances defined using the full expressive power of first order logic The classifier does not provide any additional inferences but allows PowerLoom to eagerly pre compute and cache subsumption relationships which can then be utilized over and over without having to re derive them PowerLoom also provides some experimental abductive and partial match reasoning to handle incomplete knowledge bases e Meta representation and reasoning Concepts relations contexts rules queries etc are all first class citizens in the domain of discourse Therefore they can have assertions made about them as well as reasoned about This mechanism is commonly used by the system itself e g to assert that a relation is single valued or transitive that a concept is closed etc e Explanation PowerLoom can explain its reasoning by recording inference trees and then rendering those into human understandable explanations PowerLoom also has an experimental WhyNot facility to explain inference failures where no successful proof tree was found e Contexts and modules Contexts and modules provide separate name and assertion spaces with inheritance which implement a powerful structuring mechanism for KBs Contexts allow encapsulation and organization of knowledge efficient inference by separating irrelevant knowledge or by separating ontologies and assertion spaces from volatile inference worlds truth maintenance via inference cache
176. ng from one module to another we are speaking literally a change module com mand or one of its equivalents is invoked to switch from one module to another PowerLoom comes with some modules already built in The module named PL KERNEL contains a set of general purpose concept and relation definitions that collectively form the foundation for constructing application specific knowledge bases PowerLoom attaches specialized reasoners to many of the relations in PL KERNEL The command interpreter starts up in a module named PL USER That module is initially empty and is intended as a convenient place to experiment with PowerLoom a Many of the Powerloom API procedures take a module argument that causes a temporary switch to a different module within the scope of that procedure Chapter 3 Annotated Example 11 3 Annotated Example The section presents a small example of a PowerLoom knowledge base It introduces the fundamental PowerLoom modelling concepts and illustrates the syntax of basic PowerLoom declarations assertions and commands This section can be read stand alone but readers who intend to use PowerLoom to create their own models are encouraged to load the demo file and run the examples live The conceptual terms introduced in this section include modules concepts relations functions instances propositions assertions queries retraction positive and negative facts clipping rules and contexts 3 1 Us
177. nk of the module is retrieved from the server and cached in the client Hence when the user expands a tree in the browser the GUI will not need to retrieve more Chapter 8 PowerLoom GUI 87 data from the server Also the GUI takes full advantage of Swing s ability to control the application via keyboard input For example to create a new instance named newName a user needs to simply type the key sequence Ctrl I newName RET Context Sensitivity For any given object that is displayed in a user interface there is a set of actions that can be performed on that object Additionally the actions that can be performed on an object depend on where the object is displayed Therefore the GUI pervasively supports context sensitive popup menus When users right click on an object a list of appropriate actions will be presented in a menu For example when a user right clicks on a concept in the Knowledge Browser the list of possible actions shown in Figure 8 1 is presented 4 Concepts for AIRCRAFT KB Relations for AIRCRAFT KB CONCEPT ROOT 5 RELATION ROOT C ACTIVE GUIDANCE TYPE R ATTACKED BY 9 FNTITY F CAIIRFR 9 PHYSICAL ENTITY R CARRIED BY 9 ENCINE F CEILING C DIESEL ENGINE r COST C GAS TURDINC CMGINC F CRCW ROLCS GASOLINE EMC JET ENGINE Edit GASOLINE ENGINE Ctrl C t CL et Spese Add subconcept of GASOLINE ENGINE t bA i x STEAM ENG Create instante of GASOLINE ENGINE E
178. nstance name completion is available while typing instance names by typing Ctrl right A user may choose to abort the edited extension by clicking the Cancel button If the user clicks the Commit button the relation s extension will be updated by asserting and retracting appropriate propositions Chapter 8 PowerLoom GUI 97 8 4 8 Ask and Retrieve Queries Figure 8 7 Query Dialog The Query dialog shown in Figure 8 7 can be opened by selecting the Query Query menu item by typing Ctrl Q or by pressing the query toolbar button The Query dialog consists of a text area for typing the query a results table for displaying the results of the query a query list for saving and selecting pre saved queries and an options subpane for configuring various query parameters The query input pane supports features similar to that of the proposition editor includ ing Emacs key bindings parenthesis matching and completion Queries can be executed Chapter 8 PowerLoom GUI 98 by typing Ctrl RET or by clicking on the Execute button at the bottom of the dialog After a query has executed results will be displayed in the results table or a No results found indicator will flash The column headers for the results will display the correspond ing free variables in the query Results may be sorted by clicking on a column header Doing so will sort the results by using the clicked column as an index Users may toggle ascending descending sort o
179. nt truth value self TRUTH VALUE BOOLEAN Function Return TRUE if self represents INCONSISTENT inconsistent self PROPOSITION BOOLEAN Function Return true if self is inconsistent true and false insert at self key LIKE ANY KEY SELF Method on QUERY SOLUTION TABLE value LIKE ANY VALUE SELF Insert value identified by key into self If a solution with that key already exists destructively modify it with the slot values of value This is necessary to preserve the order of solutions in self invert truth value self TRUTH VALUE TRUTH VALUE Function Return the logical negation of self known truth value self TRUTH VALUE BOOLEAN Function Return TRUE if self is a known truth value that is either TRUE or FALSE but not UNKNOWN INCONSISTENT etc length self INTEGER Method on QUERY SOLUTION TABLE Return the number of entries in self list features LIST Command Return a list containing two lists a list of currently enabled PowerLoom features and a list of all available PowerLoom features list unclassified instances module NAME N Command local BOOLEAN CONS OF LOGIC OBJECT Collect all instances in module or in any module if module is NULL that were not or will not be classified due to their lack of non inferable primitive type assertions list unclassified relations module NAME N Command local BOOLEAN CONS OF NAMED DESCRIPTION Collect all named des
180. nuity between operations For example if a user opens a browser and browses the AIRCRAFT KB and then opens a query dialog it makes sense for the query dialog to use the AIRCRAFT KB module instead of some other module The rightmost section of the status bar contains messages that pertain to the current state of the application For example if a user selects a concept and then clicks the cut toolbar button a message will appear in the rightmost status bar prompting the user to select another concept and perform a paste action The status bar may be hidden and unhidden by selecting the View Statusbar menu item Figure 8 2 shows a few internal frames that are open The function of each frame is identified in the frame s title bar and each type of frame has a unique icon in it s upper left hand corner In this example the three open frames are used to browse the KB query the KB and edit an instance A user typically follows a workflow cycle similar to the following sequence 1 The user launches the GUI by executing the powerloom script clicking on a desktop icon or on a hyperlink in a browser 2 The GUI is loaded on the user s machine If the GUI was launched via Java Web Start not yet fully supported the entire application may need to be downloaded or updated before execution begins 3 The GUI reads a preferences file stored in a default location on the user s local machine If this is the first time the application is being execu
181. o it A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s evaluate command STRING module name STRING Function environment ENVIRONMENT OBJECT Evaluate the command represented by the string command within module and return the result Currently only the evaluation of possibly nested commands and global variables is supported A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get arity relation name STRING module name STRING Function environment ENVIRONMENT INTEGER Return the arity of the relation named relation name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get child modules name STRING environment ENVIRONMENT Function PL ITERATOR OF MODULE Return the modules that are immediate children of module name If no module named name exists a Stella no such context exception is thrown s get concept name STRING module name STRING Function environment ENVIRONMENT LOGIC OBJECT Return a class concept named name that is local to or visible from the module module name A module name of null or the empty string refers
182. objects recorded in self record justifications BOOLEAN Function Return TRUE if every query records justifications to enable the explanation of con cluded results Chapter 11 Miscellaneous 125 register computation function name STRING Function code FUNCTION CODE arity INTEGER Creates a registration entry for name as a computation which executes code Essen tially just builds the Stella meta information tructure needed to funcall name as a computation function by the computation specialist The function definition in code needs to accept ARITY Stella OBJECTs as arguments and return a Stella OBJECT suitable for PowerLoom use These are generally LOGIC OBJECTs and the literal wrappers FLOAT WRAPPER INTEGER WRAPPER and STRING WRAPPER register computation function name stella name STRING Command native name STRING arity INTEGER registers a computation function stella name based on the native name for the par ticular programming language in question Use of this command makes the resulting code or knowledge bases non portable to other target languages register logic dialect print function dialect KEYWORD Function fn FUNCTION CODE WRAPPER Register fn as a logic object print function for dialect Each function should have the signature self OBJECT stream OUTPUT STREAM Any return values will be ignored register specialist function name STRING Function code FUNCTION CODE Creates
183. odules name STRING environment ENVIRONMENT Function PL ITERATOR OF MODULE Return the modules that are immediate parents of module name If no module named name exists a Stella no such context exception is thrown s get proposition relation and arguments STRING Function module name STRING environment ENVIRONMENT PROPOSITION Return a proposition matching relation and arguments that has been asserted or inferred by forward chaining relation and arguments is a string that begins with Chapter 6 PowerLoom API 56 a left parenthesis followed by a relation name one or more argument identifiers and terminated by a right parenthesis Each argument identifier can be the name of a logical constant a literal reference e g a number the null identifier or a variable an identifier that begins with a question mark Each occurrence of a null or a variable acts like a wild card If more than one proposition matches the input criteria the selection among satisficing propositions is arbitrary This procedure is normally applied to single valued relations or functions A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get propositions relation and arguments STRING Function module name STRING environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return propositions matching relation and
184. of iterators that have more than one return variable For convenience they also work with index 0 on the value of PlIterator objects that have only a single return variable For example Get values of object name age in the iterator and then print out a message with the name and age position 1 amp 2 This skips the PowerLoom object bound to X in position O Pliterator iter PLI sRetrieve all and name x name age x age null null Chapter 6 PowerLoom API 67 while iter nextP 1 System out println PLI getNthString iter 1 null null is PLI getNthInteger iter 2 null null years old 6 2 3 10 Utility Classes for Java To make interoperation of Stella and Java easier there are several convenience classes for wrapping Stella iterators and having them behave like Java enumerations or iterators These convenience classes are in the edu isi stella javalib package ConsEnumeration java Enumeration class for Cons objects ConsIterator java Iterator class for Cons objects StellaEnumeration java Enumeration interface to Stella s Iterator StellaIterator java Iterator interface to Stella s Iterator All of the iterators and enumerators return objects that are actually of type Stella_ Object but the signature specifies java lang Object as required for compatibility with the standard Java signature The Cons classes take a Cons in their constructor The Stella classes take a edu is
185. of the relation gt Its not necessary that we exactly specify the number of employees in a company Below all we know about ZZ Productions is that they have fewer than 20 employees assert and company zz productions lt number of employees zz productions 20 Chapter 3 Annotated Example 15 These facts are sufficient to classify ZZ Productions as a small business retrieve all small company x gt There are 2 solutions 1 X ZZ PRODUCTIONS 2 X ACME CLEANERS 3 7 Negation and Open and Closed World Semantics PowerLoom implements a three valued logic the truth value of each proposition entered into a PowerLoom knowledge base is recorded as being either true false or unknown Many other systems e g relational DBMSs implement a two valued logic wherein if a fact is not asserted to be true it is assumed to be false The PowerLoom command ask returns one of three five values true if it can prove the truth of a proposition false if it can easily prove the falsity of a proposition and otherwise it returns unknown The values default true and default false are also possible if defaults are used Below PowerLoom knows nothing about a newly introduced concept s corporation so ask returns unknown to both a positive query and its negation defconcept s corporation c corporation ask s corporation zz productions gt UNKNOWN ask not s corporation zz productions UNKNOWN If
186. on relation and arguments OBJECT module MODULE Function environment ENVIRONMENT PROPOSITION Return a proposition matching relation and arguments that has been asserted or in ferred by forward chaining relation and arguments is a sequence containing objects and nulls The first argument must be the name of a relation A null value acts like a wild card If more than one proposition matches the input criteria the selection among satisficing propositions is arbitrary This procedure is normally applied to single valued relations or functions get propositions relation and arguments OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return propositions matching relation and arguments that have been asserted or in ferred by forward chaining relation and arguments is a sequence containing objects and nulls The first argument must be the name of a relation A null value acts like a wild card get propositions in module module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return propositions that have been conceived in the module module get propositions of object LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return all propositions that have object among their arguments and that are TRUE in the scope of the module module get range relation LOGIC OBJECT LOGIC OBJECT Function Return the type
187. on that takes arguments of type number The requirement that relations be defined before they are referenced can be inconvenient at times For example suppose we wish to define parent as a person who is the parent of another person and we also wish to state that the first argument to the parent of relation has type parent defconcept parent p lt gt and person p exists c parent of p c defrelation parent of p parent c person In this example the first reference to parent of occurs before it is defined PowerLoom permits circular references such as these as long as they occur within definitions It does so by deferring evaluation of rules that occur within definitions Here is a specification that is logically equivalent but is not legal because the parent of relation appears in an assertion before it is defined defconcept parent p assert forall p lt gt parent p and person p exists c parent of p c defrelation parent of p parent c person So when does the rule inside of the first parent definition get defined All axioms facts and rules that appear within the boundaries of a definition are evaluated just prior to the next occurrence of a PowerLoom query Hence in the example above where the rule occurred within the definition there was no error because evaluation of that rule occured sometime after the second definition which defines the otherwise problematic ref
188. or shown in Figure 8 5 consists of a text field for entering the proposition and a set of buttons for performing actions on the proposition The buttons allow a user to assert deny or retract the typed proposition There are several text based facilities which support efficient editing of propositions First the editor supports many Emacs style keybindings which facilitate editing of Lisp like expressions including selecting entire parenthesis delimited subexpressions jumping backward and forward over subexpressions and navigating up and down expression trees P Edit Proposition vu DCFORALL v 7e lt NOT ENGINE TYPE v e AND UNPOWERED VEHICLE 7w NOT CENGINE COUNT v 022222 Module AIRCRAFT KB v Cancel Figure 8 5 Proposition Editor In addition to Emacs keybindings the proposition editor has a matching parenthesis highlighter When the cursor is placed before a left parenthesis the matching right paren thesis is highlighted and when the cursor is placed after a right parenthesis the matching left parenthesis is highlighted The proposition editor also supports symbol completion The GUI uses a predictive backtracking parser to analyze partial input of propositions Based on the analysis the parser is able to recommend appropriate completions For example if the user types f and then selects the completion action by typing Ctrl right the parser will recommend a list of completions including the f
189. or FALSE propositions attached to self all facts of n n INTEGER amp rest instanceRefs NAME CONS N Command OF PROPOSITION This is a generalization of all facts of which see With n 0 and only one instance this command behaves just like all facts of Otherwise returns a cons list of all definite TRUE or FALSE propositions that reference any of the instances listed in instanceRefs plus if n gt 1 all propositions that reference any instances that are arguments of propositions collected in the previous step plus if n gt 2 and so on That is if we only consider binary propositions this can be viewed as growing a graph with instances as its nodes and predicates as its arcs starting from the set of seed instanceRefs to depth n 1 Caution with a fully connected KB and large enough n this could return the whole knowledge base The returned propositions include those asserted to be true or false by default but it does not include propositions that are found to be true only by running the query engine Facts inferred to be true by the forward chainer will be included Hence the returned list of facts may be longer in a context where the forward chainer has been run then in one where it has not see run forward rules all inconsistent propositions module MODULE Function local BOOLEAN ITERATOR OF PROPOSITION Iterate over all conceived propositions visible from module that have an inconsis tent truth value If loc
190. or failure of a proof select test result success BOOLEAN terminal BOOLEAN Function frame CONTROL FRAME KEYWORD Helping function for specialists testing the validity of a fully bound inference frame Based on the test result success and reversePolarity set the truth value of frame and return an appropriate keyword The keyword will be either final success terminal failure if terminal is true Otherwise it will be final success or failure set error print cycle iINTEGER Command Set number of cycles between which error rates are saved to the file established by the last call to save all neural networks appended with extension err A number 0 or NULL turns off periodic saving set inference level level NAME module NAME KEYWORD N Command Set the inference level of module to the level specified by levelKeyword If module is NULL set the level globally set marker self object OBJECT Method on MARKER TABLE Record membership of object in the marker storage object self set num neighbors d INTEGER Command Sets the number of nearest neighbors to predict from set num training per case d INTEGER Command Sets the number of training examples for each case in the training set set powerloom feature feature KEYWORD Function Enable the PowerLoom environment feature feature set save network cycle iINTEGER Command Set number of cycles between which networks are sav
191. orall symbol and all concepts and relations that begin with the letter f 8 4 6 Choosers In a number of situations an object of a specific type must be selected For example when selecting a superconcept in a concept editor the user should be presented with a list of Chapter 8 PowerLoom GUI 96 existing concepts In these cases a chooser dialog is presented to the user which displays a filterable list of candidate objects As the user types a name of the object in the name text field the list of objects is filtered so that only objects which begin with the typed prefix are displayed Choosers are available for modules concept instances and relations A variable chooser allows the user to type a variable name and select a type from a concept from a list E Edit Extension Relation wingspan Module AIRCRAFT KB v Results ies o a Figure 8 6 Extension Editor 8 4 7 Extension Editor The extension editor shown in Figure 8 6 allows editing of a concept or relation s extension It can be opened by right clicking on a concept or relation in the browser or by selecting the Query gt Edit Extension menu item The extension editor presents a relation s extension as a list of tuples in table format The user may add new tuples by typing names of instances at the bottom of the table and may alter existing tuples by double clicking on a table cell and typing in a new value I
192. owerLoom gives programmers the option to explicitly declare that concept or a relation operates under the assumption of closed world semantics See also Open World Semantics Concept A concept defines a category or class of individuals PowerLoom categorizes a concept as a special kind of relation The distinction between a concept and a unary relation is subtle some logicians do not believe that there is any distinction In linguistics the distinction is that between a noun and an adjective In logic the test we favor is whether or not the relation has a domain a unary relation has a domain while a concept does not For example the relation married has domain person while the concept married person does not have a domain or is its own domain Constraint Constraint at its most general is a synonym for rule Often a constraint is conceptualized as a rule that restricts the types of the arguments that can appear within a tuple Context Default Rule A default rule expresses an conditional implication that applies only when its consequent is consistent with current state of the knowledge base In other words the rule applies only when it will not lead to a contradition Definition A definition binds a name to a logical expression PowerLoom syntax defines several operators with names of the form defxxx e g defconcept and defrule that declare definitions for various types of entities 1
193. own But if we assert that the relation works for is closed then PowerLoom will assume that Jerome only works for ZZ Productions ask not works for jerome megasoft UNKNOWN assert closed works for ask not works for jerome megasoft TRUE The reasoning employed to achieve the above result that Jerome does not work for MegaSoft is called negation as failure which means that if a proof of a proposition fails then one may assume that the proposition is false We can achieve a negation as failure result a second way i e other than by using a closed world assumption by employing the query operator fail Here we retract the closure assumption for works for and achieve the desired result using fail retract closed works for Chapter 3 Annotated Example I ask not works for jerome megasoft UNKNOWN ask fail works for jerome megasoft TRUE When you see the operator not in an SQL query or a Prolog program it really stands for fail 3 8 Retraction Below we introduce a few new terms for defining geographic information We define a relation called contains to assert that one geographic location the second argument to contains is located within another defconcept geographic location defconcept country 1 geographic location defconcept state 1 geographic location defconcept city 1 geographic location defrelation contains 11 geographic location
194. p time you might want to create a Lisp image that has all of PowerLoom preloaded Now type in package STELLA to enter the STELLA Lisp package where all the PowerLoom code resides Alternatively you can type Chapter 10 Installation 106 powerloom which will bring up a PowerLoom listener that will allow you to execute PowerLoom commands IMPORTANT All unqualified Lisp symbols in this document are assumed to be in the STELLA Lisp package Moreover the STELLA package does NOT inherit anything from the COMMON LISP package see the file sources stella cl lib cl setup lisp for the few exceptions Hence you have to explicitly qualify every Lisp symbol you want to use with CL For example to get the result of the previous evaluation you have to type CL instead of 10 4 C Installation To compile the C version of PowerLoom change to the native C directory of Power Loom and run make like this cd native cpp powerloom make This will compile all PowerLoom and STELLA files as well as the C garbage collector It will then generate static or dynamic libstella liblogic and other library files in the directory native cpp 1lib which can be linked with your own C translated PowerLoom or other code To test whether the compilation was successful you can run PowerLoom from the top level PowerLoom directory using the powerloom script or powerloom bat under a Windows command prompt
195. persand gt Special variable marker The character mappings use uppercase characters if the basic identifier uses mixed or lower case The mappings use lowercase characters if the basic identifier uses upper case Stella modules are mapped to Java packages The basic system distribution includes the following package hierarchy edu isi stella javalib powerloom logic pl_kernel_kb loom_api Basic system functionality and data structures such as Cons and List objects are defined in stella PowerLoom s logic concepts relations rules etc are defined in the logic package There is a set of interface functions in the PLI class in the powerloom package They are described in their own section below We recommend the following import statements in Java files that use PowerLoom import edu isi stella import edu isi stella javalib import edu isi powerloom PLI import edu isi powerloom logic Functions in Java terms static Methods are translated as static methods on the class of their first argument as long as that argument is not a primitive type and is in the Chapter 6 PowerLoom API 62 same Stella module Functions which take no arguments those whose first argument is a primitive type and those whose first argument is a class not defined in the same module are all placed into a class with the same name as the Stella module in which it appers It will be in the package corresponding to that Stella module J
196. ple output variables each element generated by the returned iterator will be a list of variable bindings that can be accessed using the various get nth functions The list of output variables does not need to be declared in which case they are taken to be the open variables in the query proposition in the order in which they were encountered If order does matter or should be different from its default it can be forced by declaring the set of output variables run forward rules module OBJECT force BOOLEAN Function Run forward inference rules in module module If module is NULL the current module will be used If forward inferencing is already up to date in the designated module no additional inferencing will occur unless force is set to TRUE in which case all forward rules are run or rerun Calling run forward rules temporarily puts the module into a mode where future assertional monotonic updates will trigger additional forward inference Once a non monotonic update is performed i e a retraction or clipping of relation value all cached forward inferences will be discarded and forward inferencing will be disabled until this function is called again s ask query STRING module name STRING Function environment ENVIRONMENT TRUTH VALUE Returns a truth value for query in module module name and environment query has the same syntax as the PowerLoom ask command which see but with the ask operator omitted Different
197. possible to use a version in which Stella and PowerLoom objects are implemented using Common Lisp structs instead This is controlled by the special variable cl user load cl struct stella If this is set to cl t then the struct version will be loaded This needs to be set before loading the load powerloom 1lisp file 6 2 2 C API lt to be written gt 6 2 3 Java API This section tells how to call the API functions in PowerLoom s Java implementation from a Java program The Java translation is written for Java version 1 2 All of the PowerLoom Interface functions appear as static methods of the class edu isi powerloom PLI 6 2 3 1 Initializing PowerLoom PowerLoom needs to run initialization functions to set up its environment for proper op eration when it starts up The simplest method for initializing PowerLoom is to use the static method call PLI initialize This must be called before using any PowerLoom features and before loading any Pow erLoom knowledge bases It may be called more than once without ill effect 6 2 3 2 PowerLoom Java Conventions PowerLoom s Java code is automatically generated by a translator from underlying Stella code The character set for legal Stella names is larger than the character set for legal Java identifiers so there is some mapping involved PowerLoom names are words separated by hyphen characters For Java we have attempted to closely follow the Java conventions e Clas
198. powerloom c Running C version of PowerLoom Initializing STELLA Initializing PowerLoom Welcome to PowerLoom 4 0 0 Copyright C USC Information Sciences Institute 1997 2010 PowerLoom comes with ABSOLUTELY NO WARRANTY Type copyright for detailed copyright information Type help for a list of available commands Type demo for a list of example applications Type bye exit halt quit or stop to exit PL USER This will run various PowerLoom startup code and then bring up a PowerLoom command loop where you can execute commands The c argument tells the script to run the C version of PowerLoom which is also run by default as long as the C version was compiled If the C version was not compiled or the java argument was given instead the Java version of PowerLoom will be run Type demo to bring up a menu of available demos type Chapter 10 Installation 107 run powerloom tests to run the PowerLoom test suite or type exit to exit PowerLoom 10 5 Java Installation Nothing needs to be done to install the Java version Because Java class files are platform independent they are already shipped with the PowerLoom distribution and can be found in the directory native java and its subdirectories Additionally they have been collected into the file native java lib powerloom jar in the PowerLoom directory To try out the Java version o
199. proof structure literally i e no uninteresting nodes such as patterns or AND introductions are suppressed consify self CONS Method on QUERY ITERATOR Generate all solutions for the query self and collect them into a cons list of result tuples If SINGLETONS TRUE collect a list of atoms rather than a list of lists for tuples of arity 1 consify self CONS Method on QUERY SOLUTION TABLE Collect all solutions of self into a cons list and return the result consify current solutions self CONS Method on QUERY ITERATOR Collect the current solutions of self into a cons list of result tuples If SINGLETONS TRUE collect a list of atoms rather than a list of lists for tuples of arity 1 Chapter 11 Miscellaneous 115 consify justification self JUSTIFICATION style KEYWORD Function CONS Return a CONS tree representation of the proof self Each proof step is represented as a CONS tree of the form lt proposition gt lt key gt lt value gt lt antecedent gt where each lt antecedent gt is a CONS tree representing a subproof style indicates what nodes in the proof tree should be suppressed RAW preserves the original structure literally VERBOSE keeps AND introductions but suppresses all auxiliary non logical nodes such as pattern nodes and BRIEF additionally suppresses AND introduction nodes constant objectRef OBJECT BOOLEAN Function Return TRUE if objectRef denotes a literal or scalar
200. pter 3 Annotated Example 19 assert headquartered in megasoft dallas retrieve all headquartered in megasoft x There is 1 solution 1 X DALLAS 3 10 Rule based Inference Suppose that we want to retrieve all geographic locations that are contained in the Southern US based on the set of assertions about geography that we entered in earlier The following query returns only one of such location retrieve all contains southern us x There is 1 solution 41 X TEXAS We would like the cities Austin and Dallas to be retrieved as well To do this we can assert a rule that states that contains is a transitive relation defrule transitive contains gt and contains 11 712 contains 12 7 13 contains 11 13 The defrule declaration does two things it asserts a proposition and it associates a name with that proposition in the above case the name is transitive contains This name is used by the system in displaying traces of its inferencing It also makes redefinition of the proposition easier If we wish to retract an unnamed proposition it is necessary to explicitly retract that proposition using a syntax identical to the assertion If on the other hand a proposition has a name then a new defrule declaration that uses the same name will automatically retract any existing proposition having the same name Our transitive closure rule failed to include any logical quantifiers for the variables 11
201. queries upon next invocation retract facts of instance self LOGIC OBJECT Function Retract all definite TRUE or FALSE propositions attached to self retrieve partial amp rest tree PARSE TREE QUERY ITERATOR N Command Partial match version of retrieve which see that generates scored partial solutions based on the current partial match strategy By supplying BEST instead of ALL or by adding the option SORT BY SCORE the generated solutions will be sorted so that solutions with higher scores come first Use the MATCH MODE option to override the global default setting established by set partial match mode e g use MATCH MODE NN to use the neural net partial match mode The MINIMUM SCORE option can be used to only retrieve solutions that have at least the specified minimum match score By default retrieve partial does not maximize the match scores of its returned bindings To only get maximal scores use MAXIMIZE SCORE TRUE this is not yet implemented you can use ask partial to maximize scores for individual solutions by hand run forward rules moduleRef NAME amp rest force KEYWORD N Command Run forward inference rules in module moduleRef If moduleRef is NULL the current module will be used If forward inferencing is already up to date in the designated module no additional inferencing will occur unless the optional keyword force is included in which case all forward rules are run or rerun Calling r
202. r feature may change get arity relation LOGIC OBJECT INTEGER Function Return the arity of the relation relation get binary proposition relation LOGIC OBJECT arg1 OBJECT Function arg2 OBJECT module MODULE environment ENVIRONMENT PROPOSITION Return a proposition such that relation arg1 arg2 is true The relation argument must be bound to a relation One or both of the arg and arg2 arguments may be set to NULL which is interpreted as a wildcard If more than one proposition matches the input criteria the selection is arbitrary This procedure is normally applied to single valued relations or functions get binary propositions relation LOGIC OBJECT arg1 OBJECT Function arg2 OBJECT module MODULE environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return propositions such that relation arg1 arg2 is true The relation argument must be bound to a relation One or both of the arg and arg2 arguments may be set to NULL which is interpreted as a wildcard get child modules module MODULE PL ITERATOR OF MODULE Function Return the modules that are immediate children of module get column count obj OBJECT INTEGER Function Return the number of columns in obj which must be of type proposition cons vector or PL iterator For a proposition the number includes both the predidate and arguments For the PL iterator case the number of columns is for the current value of the iterator
203. r lists functions that collectively define the PowerLoom API The first section describes the API functions themselves The signature is the basic Stella signature In formation on how to translate the names of the functions and their arguments into the programming languages Common Lisp C or Java is given in the Language Specific In terface section 6 1 API Functions Many of the functions take a module argument that causes the function to be evaluated in the context of that module Passing in a NULL value for the module argument means that evaluation takes place in the current module The module argument is frequently followed by an environment argument that specifies which inference environment should be assumed during evaluation Values for environment are ASSERTION ENV TAXONOMIC ENV and INFERENCE ENV ASSERTION ENV specifies that a knowledge base query or lookup should take into account only explicitly asserted propositions TAXONOMIC ENV specifies that a knowledge base query should take into account explicitly asserted propositions plus any rules that specify subsumption relationships INFERENCE ENV specifies that a knowledge base query should take all relevant propositions into account including those generated during forward inferencing A NULL value for the environment argument defaults to TAXONOMIC ENV Many of the functions that take PowerLoom or Stella objects as inputs al
204. r translated versions and as many or more compiled versions thereof The actual PowerLoom libraries that you have to ship with an application however are quite small For example the Java jar files powerloom jar and stella jar are only 2 3 MB total 4 MB including Java sources The dynamic C libraries libstella so and liblogic so compiled on a Linux platform are about 8 MB total Additionally if you don t need all the different translations of PowerLoom you can delete some of the versions to keep your development tree smaller See Section 10 6 Removing Unneeded Files page 107 To run the Lisp version of PowerLoom you will need an ANSI Common Lisp or at least one that supports CLOS and logical pathnames We have successfully tested PowerLoom with Allegro CL 4 2 4 3 5 0 6 0 and 7 0 Macintosh MCL 3 0 4 0 and 5 1 OpenMCL 1 0 SBCL 0 9 4 CMUCL 19c LispWorks 4 4 6 CLisp 2 37 Lucid CL 4 1 plus the necessary ANSI extensions and Mark Kantrowitz s logical pathnames implementation and various others Our main development platform is Allegro CL running under Sun Solaris and Linux The closer your environment is to ours the higher are the chances that everything will work right out of the box Lisp development under Windows should also be no problem To run the C version of PowerLoom you will need a C compiler such as g that supports templates and exception handling We have successfully compiled and run
205. rLoom also has a powerful relational database interface that allows it to utilize the power of databases for handling large assertion bases soon to be released One application of this PowerLoom RDBMS integration is used with ISI s Omega ontology It is also a crucial part of our KOJAK Link Discovery System e Tools and APIs PowerLoom has a host of associated tools and APIs not all of which have been released yet It comes with an interactive command line interface which is useful for developing ontologies and knowledge bases an extensive programmatic interface called PLI with Lisp C and Java bindings and a Lisp based Loom API to load legacy Loom KBs Starting with PowerLoom version 4 0 Ontosaurus and a Java based GUI have been released as part of PowerLoom Ontosaurus is a Web based KB browser that dynamically generates HTML pages viewable in a standard Web browser The Java based GUI provides a browse edit query environment for developing KBs The GUI uses a client server architecture and can be used embedded or standalone against a PowerLoom server that might be hosted among other options in a Web server such as Tomcat A soon to be released Protege plug in allows export of Protege ontologies into PowerLoom format OntoMorph is a translation system that supports writing of KB translators and importers e g to import ontologies written in other languages for example Flogic e Initial Semantic Web support Given PowerLoom s emphasis
206. rder by clicking the header multiple times If the query contains no free variables it is effectively a true false ASK operation as opposed to a RETRIEVE In this case the result will be a truth value and the column header will be labeled TRUTH VALUE If the query is the result of a partial retrieve operation an additional column containing the match score will be displayed If the user clicks on a cell in the results table the topmost browser will be updated to display the selected item Right clicking on a query result brings up a context menu which currently only contains an Explain result menu item Selecting this will present an HTML explanation in a separate window The displayed explanation may contain hyperlinked objects Clicking on a hyperlinked object will update the topmost browser to display the object Users may save frequently executed queries in a query list by clicking the Save button at the top of the options panel which will prompt them for a query name Saved queries will be stored in the preferences file using an XML representation Saved queries are stored in the combobox to the left of the save button Selection of a saved query will prefill the query dialog with the query and all saved parameters Most important PowerLoom query options are controllable in the options dialog such as number of results to retrieve inference control options such as inference level timeout moveout and various others 8 4 9 Search Use
207. red has not yet been created it is automatically created as a side effect of calling assert assert from query query CONS amp rest options OBJECT Cons N Command OF PROPOSITION Evaluate query instantiate the query proposition for each generated solution and assert the resulting propositions The accepted syntax is as follows assert from query lt query command gt relation lt relation name gt pattern lt description term gt module lt module name gt record justifications TRUE FALSE lt query command gt has to be a strict or partial retrieval command If a relation op tion is supplied lt relation name gt is used as the relation of the resulting propositions Chapter 5 Commands 29 In this case the bindings of each solution will become arguments to the specified rela tion in the order of querys output variables the arities have to match The pattern option is a generalization of this mechanism that specifies an arbitrary proposition pattern to be instantiated by the query s solution In this case lt description term gt has to be a SETOFALL or KAPPA expression whose IO variables will be bound in sequence to the bindings of a query solution to generate the resulting proposition Fi nally if a module option is specified the assertions will be generated in that module Note that for this to work the relations referenced in the query proposition or pattern have to be visible in the module Also
208. relative to which commands are interpreted For example type the demo command to see a menu of available demos Step through one or more of them to get familiar with PowerLoom Starting with version 4 0 PowerLoom also ships with an experimental PowerLoom GUI and the Ontosaurus browser which can additionally be used to edit and browse knowledge bases 1 3 1 Command Line Options There are a few command line options that can be supplied to the powerloom script powerloom c javal guil gui only e eval STELLA EXPRESSION batch help more options The first optional argument determines what version to run if both C and Java versions are installed If no specific version is specified the C version will be run if it is installed otherwise the Java version will be run see Chapter 8 PowerLoom GUI page 84 for a description of the GUI options eval STELLA EXPRESSION e STELLA EXPRESSION Specifies a STELLA expression that should be run just before the PowerLoom command loop gets initialized This expression has to be a known command such as the various PowerLoom commands since the STELLA evaluator can not yet evaluate arbitrary STELLA code For example powerloom e demo equations FALSE will run a particular demo before anything else You will need to appropriately quote special characters interpreted by the shell or the Command Prompt window batch Runs PowerLoom in batch
209. rence level set to subsumption Equivalent to query prop inference level subsumption but more efficient Chapter 7 Built In Relations 82 SUM COLLECTION sum NUMBER Relation Binds sum to the sum of the numbers in the list 1 SUPERRELATION r RELATION super RELATION Relation True iff super is a superrelation of r written in set notation super gt r This relation will generate bindings for at most one unbound argument SYMMETRIC r RELATION Relation A binary relation r is symmetric if it is commutative SYNONYM term THING synonym THING Relation Assert that synonym is a synonym of term This causes all references to synonym to be interpreted as references to term Retraction eliminates a synonym relation THING STANDARD OBJECT DYNAMIC SLOTS MIXIN Class Defines a class that must be inherited by any class that participates in the PowerLoom side of things surrogate value inverse SURROGATE Slot of Not documented THIRD ELEMENT l COLLECTION gt e THING Function Return the third element e of 71 TOTAL r FUNCTION Relation True if the function r is defined for all combinations of inputs By default functions are not assumed to be total unlike Prolog which does make such an assumption For example if we define a two argument function foo and then retrieve its value applied to some random instances a and b we get nothing back deffunction foo x y
210. rent of p1 person p2 person The assertion in this latter sequence is logically equivalent to the axiom introduced by the gt keyword in the former sequence However at the end of this sequence the parent of implies relative of rule is still in effect since it appeared on its own outside of a definition 2 3 Truth Values A PowerLoom proposition is tagged with a truth value that has one of five different settings true false default true default false or unknown The most common setting is true when we make an assertion as in assert Person Bill the propo sition Person Bill is assigned the truth value true To assign the value false to a proposition one asserts that it is not true e g assert not crook Richard The command presume is used to assign a proposition the value default true as in presume weather in Los Angeles Sunny Presuming a negated proposition assigns it the value default false The assignment of a truth value to a proposition via assert or presume can upgrade the strength of a proposition but it cannot downgrade it Hence if a proposition currently has the value unknown then it may be assigned any of the other four values If the value is default true or default false an assertion that assigns the value true or false will overwrite the existing value However if the truth value of a proposition is either true or false assigning it the value default true or default false will hav
211. ributes for program objects just by overwriting previous values PowerLoom implements a similar semantics for the special case of functions and single valued relations When a second value is asserted for one of these relations the previous value is automatically retracted We call this clipping To illustrate this behavior for both kinds of relations a function is considered a kind of relation we will define a mapping from a company to a city that contains its headquarters in two different ways deffunction headquarters c company gt city city defrelation headquartered in c company city city axioms single valued headquartered in The clause axioms single valued headquartered in tells PowerLoom that the headquartered in relation is single valued i e that it can map a company to at most one city This makes its behavior similar to that of the function headquarters Here is an example of clipping for the function headquarters assert headquarters zz productions atlanta retrieve all x headquarters zz productions gt There is 1 solution 1 X ATLANTA assert headquarters zz productions dallas retrieve all x headquarters zz productions There is 1 solution 1 X DALLAS Here is the same kind of clipping using a single valued relation assert headquartered in megasoft atlanta retrieve all headquartered in megasoft x There is 1 solution 1 X ATLANTA Cha
212. rief definitions for terms used in the PowerLoom User s Manual and or used by the knowledge representation community It is impractical to give a logically precise definition for many of these terms because their interpretation varies quite a bit In this case the glossary attempts to indicate a range of interpretations consistent with their use in PowerLoom Assertion An assertion states that a particular proposition is True or False Backward and Forward Inference BACKWARD RULE Binary Relation A relation having two arguments arity equals two often as a mapping from one concept domain to another This is by far the most common form of relation Classifier A classifier is a type of an inference engine that implements efficient strategies for computing subsumption relations between pairs of concepts or for computing instance of relations between a concept an a set of instances PowerLoom implements a classifier that can be explicitly invoked by an application program Clipping If a function or single valued binary relation maps an instance to two or more other instances a logical contradiction a clash exists If clipping is enabled PowerLoom will automatically retract all assertions but the last that lead to a clash Clipping can be toggled on or off it is enabled by default Closed World Semantics Under closed world semantics it is assumed that if proposition P cannot be proved True then assume that P is False P
213. rm gt lt relsent gt lt constant gt lt term gt Note Zero arguments allowed in KIF3 0 for lt relsent gt lt term gt lt funconst gt is currently not allowed in PowerLoom use lt funterm gt lt term gt instead lt logsent gt not lt sentence gt and lt sentence gt or lt sentence gt gt lt sentence gt lt sentence gt lt lt sentence gt lt sentence gt gt lt sentence gt lt sentence gt sentence lt sentence gt gt gt lt sentence gt lt sentence gt lt lt sentence lt sentence gt gt gt lt sentence gt lt sentence gt lt sentence gt lt sentence gt lt quantsent gt forall lt vardecl gt lt vardecl gt lt sentence gt forall lt vardecl gt lt vardecl gt sentence lt sentence gt exists lt vardecl gt lt vardecl gt lt sentence gt Chapter 13 PowerLoom Grammar 139 13 2 4 Definitions PowerLoom supports two distinct categories of definitions relation definitions including concept and function definitions and rule definitions A relation definition introduces a new logical constant and states some facts about that constant e g who its parents are in a subsumption taxonomy A rule definitions binds a new constant to a proposition so that the constant denotes the proposition and asserts the truth of that proposition Usually the proposition a
214. rs may search for objects in the KB by entering strings which match the name of the object A search dialog as shown in Figure 8 8 can be opened by selecting the Query gt Search menu item by typing Ctrl f or by pushing a search toolbar button marked by a flashlight icon inside the browser If the user activates a search toolbar button inside a navigation pane the search dialog will be configured to search for objects associated with the type of object displayed in that pane For example pushing the search button inside the concept navigation pane will configure the search dialog to look for concept objects only Searches may be constrained in several ways First the module may be specified or the user may specify that the search should be across all modules Second the types of objects to be searched is configurable For example users may search for concepts and instances instances only etc Finally users may specify that an object s name must match the beginning or end of the search string or match the search string exactly Chapter 8 PowerLoom GUI 99 Search String co Module AIRCRAFT KB Contains vi Concepts j Starts With v Relations 1 _ Case Sensitive Ends With vi Instances gt Exactly Matches Module y AIRCRAFT KB CONCEPT AIRCRAFT KB CONCEPT AIRCRAFT KB CONCEPT AIRCRAFT KB CONCEPT AIRCRAFT KB RELATION AIRCRAFT KB RELATION AIRCRAFT KB INSTANCE AIRCRAFT KB CONCEPT AIRCRAFT KB INS
215. s a true truth value It can be true either absolutely or by default is true binary proposition relation LOGIC OBJECT arg OBJECT Function value OBJECT module MODULE environment ENVIRONMENT BOOLEAN Return TRUE if the proposition relation arg value has been asserted or inferred by forward chaining is true proposition proposition PROPOSITION module MODULE Function environment ENVIRONMENT BOOLEAN Return TRUE if proposition is TRUE in the module module is true unary proposition relation LOGIC OBJECT arg OBJECT Function module MODULE environment ENVIRONMENT BOOLEAN Return TRUE if the proposition relation arg has been asserted or inferred by forward chaining is unknown tv TRUTH VALUE BOOLEAN Function Tests whether tv is an unknown truth value iterator to pl iterator selfITERATOR PL ITERATOR Function Convert an arbitrary Stella iterator into an API iterator length self INTEGER Method on PL ITERATOR Number of items remaining in self Non destructive list to pl iterator self LiST PL ITERATOR Function Convert a Stella list into an API iterator load filename STRING environment ENVIRONMENT Function Read logic commands from the file named filename and evaluate them The file should contain an in module declaration that specifies the module within which all remaining commands are to be evaluated The remaining commands are evaluated one by one applying the
216. s ese leita ded noae pi i pe daa bee Rogue 56 S CSt VeELAtC LOM 6 has dead e eds naan salad meee 56 s get relation extension 57 S get rulesS x 2 nea EuR RENERE SENES 57 s is true propositiOon c er 5T S print r le8 s ic REP ir d 5T s register computation function 132 s register specialist function 132 s retract proposidtiOn e e e er 5T S retrieve o did dd hb uua i n tas 5T S Ssave module eli es aoc e pie peus 58 satisfies ogos see 9 e eri eer te 126 save all neural networks 126 save module ases Ra her MP 38 58 SCALAR bo se hr fert RES E n RIDERE 80 Function Index SCALAR INTERVAL 2 0 cee eee cee eee eee 80 SECOND ELEMENT x eR re bRRSR eb 80 select proof result gio etx ERR 127 select test result lors Rr e Ree 127 Set error print CyCl sss renin ni r Enori rips 127 Set feature i llb Y we EOE EN eee we YA 38 set inference level 2 sh tees deen 127 set load path v ule e E dep DRE eed 38 set marker on MARKER TABLE 127 set num neighbors sues 127 set num training per case 127 set powerloom feature sss ss 127 set save network cycle 06 127 SELOP ois die sess an m REDI ad oud en EU ERIT 80 SBALLOW QUERY resriebesekrvue ea ber aden tenes 80 SINGLE VALUED 224 2 venei RI eines ie heen ERS 80 sort on QUERY SOLUTION TABLE
217. s names begin with a capital letter and each word is capitalized The hyphens from the PowerLoom names are removed Example string wrapper gt StringWrapper Exceptions are made for class names that would otherwise conflict with normal Java Classes In that case the prefix Stella is added to each class name At the moment this applies only to the following exceptions object gt Stella Object class gt Stella Class Chapter 6 PowerLoom API 61 e Method and Function names begin with a lower case letter but each subsequent word is capitalized The hyphens from PowerLoom names are removed Example wrapper value gt wrapperValue e Storage slots are implemented as Java fields The names begin with a lower case letter but each subsequent word is capitalized The hyphens from PowerLoom names are removed Example dynamic slots gt dynamicSlots e Global and Special variable names are written in all uppercase The hyphens from PowerLoom are replaced by underscore _ characters By convention special variables are written in PowerLoom with surrounding asterisk characters The asterisks are replaced by dollar signs Example html quoted characters gt HTML_QUOTED_CHARACTERS The most common non alphanumeric characters are mapped as follows A full set of mappings is in section Section 6 2 3 7 Java Character Mapping page 65 gt P for Predicate gt X eXclamation gt B Buck R peRcent gt A Am
218. s visible from module A term can be an instance or individ ual as well as a description Only terms that haven t been deleted will be considered If local only return terms created locally in module all types self OBJECT CONS OF NAMED DESCRIPTION Function Return a set of all of the types that are satisfied by self Chapter 11 Miscellaneous 112 all unnamed terms module MODULE local BOOLEAN ITERATOR Function Iterate over all unnamed terms visible from module A term can be an instance or individual as well as a description Only terms that haven t been deleted will be considered If local only return terms created locally in module allocate supported closure iterator startnode CONS Function allocateadjacencyiterator FUNCTION CODE filterfunction FUNCTION CODE SUPPORTED CLOSURE ITERATOR Similar to allocate transitive closure iterator which see but return a SUPPORTED CLOSURE ITERATOR instead allocate transitive closure iterator startNode OBJECT Function allocateAdjacencyIterator FUNCTION CODE filterFunction FUNCTION CODE ITERATOR Return an iterator that generates the transitive closure of applying iterators gener ated by allocateAdjacencylterator to startNode If filterFunction is non null that function is applied as a filter to each node generated nodes filtered out still generate descendants but they don t get returned apply ask amp body body coNs OBJECT Macro
219. se of get nth float on PL iterators with only a single return variable If sequence cannot be turned into a floating point value an exception will be thrown get nth integer sequence OBJECT n INTEGER module MODULE Function environment ENVIRONMENT INTEGER Return an integer representation of the value in the nth column of sequence Counting starts at zero Unless n is zero sequence must be of type proposition cons vector or PL iterator A zero column number returns a proposition s relational predicate For the PL iterator case the the current value pointed to by the iterator is used If this is not an integer value then an exception will be thrown As a special case a column number of zero will also return the integer value of sequence itself if it is not one of the types enumerated above This allows the use of get nth integer on PL iterators with only a single return variable If sequence cannot be turned into an integer an exception will be thrown Chapter 6 PowerLoom API 46 get nth logic object sequence OBJECT n INTEGER Function module MODULE environment ENVIRONMENT LOGIC OBJECT Return a logic object representation of the value in the nth column of sequence Counting starts at zero Unless n is zero sequence must be of type proposition cons vector or PL iterator A zero column number returns a proposition s relational predicate For the PL iterator case the the current value pointed to by the iterator
220. set of fillers of the relation r applied to i The cardinality function returns a value only when the relations range min cardinality and range max cardinality compute identical values i e when the best lower and upper bounds on the cardinality are equal Each of these bounding functions employs a variety of rules to try and compute a tight bound RANGE CARDINALITY LOWER BOUND r RELATION i THING Relation Ib INTEGER Relation that specifies a lower bound on the cardinality of the set of fillers of the relation r applied to i The difference between range cardinality lower bound and range min cardinality is subtle but significant Suppose we state that nine is a lower bound on the number of planets in the solar system and then ask if eight is also a lower bound assert range cardinality lower bound hasPlanets SolarSystem 9 ask range cardinality lower bound hasPlanets SolarSystem 8 gt TRUE PowerLoom will return TRUE However if we ask if the minimum cardinality of the solar system s planets is eight we get back UNKNOWN Chapter 7 Built In Relations 79 ask range min cardinality hasPlanets SolarSystem 8 gt UNKNOWN because eight is not the tightest lower bound RANGE CARDINALITY UPPER BOUND r RELATION i THING Relation ub INTEGER Relation that specifies an upper bound on the cardinality of the set of fillers of the relation r applied to i see the discussion for range cardinality lower
221. side applications 8 5 Future Work There are many areas where further development and improvement is needed including Chapter 8 PowerLoom GUI 101 8 5 1 Large KBs Currently when a module is selected the GUI attempts to retrieve all concepts relations and instances that are contained in that module For large knowledge bases that might contain millions of objects and assertions this is clearly infeasible To handle such situations we need to develop more sophisticated retrieval and caching strategies to only materialize partial views of a KB in the GUI Scrolling down a list for example might then trigger the retrieval and display of additional objects on demand while older inactive objects are being flushed from the cache 8 5 2 Undo There is currently no undo facility and changes to the KB are written through to the server as soon as a user clicks OK in an editing dialog To support this properly we would need a snapshot mechanism that would allow rollback to earlier states of the KB 8 5 3 Drag and Drop Adding a drag and drop capability would make ontology editing easier than is currently possible For example one concept could be dragged on top of another to move the object from its current position We believe that the existing data transfer framework could be leveraged to implement a robust drag and drop facility 8 5 4 Scrapbook In creating and editing ontologies it is sometimes desirable to maintain heterogeneo
222. so have an analog version whose name starts with the prefix s that take strings as inputs This is provided as a convenience so that programmers will not necessarily need to manipulate PowerLoom objects directly ask query CONS module MODULE environment ENVIRONMENT Function TRUTH VALUE Returns a truth value for query in module and environment query has the same syn tax as the PowerLoom ask command which see but with the ask operator omitted For example here are some legal query arguments happy Fred happy Fred inference level assertion happy Fred inference level assertion timeout 1 0 As a convenience a query argument whose first element is a symbol is interpreted as a sentence that is queried without any options For example happy Fred is a legal query argument Note that for a sentence whose relation is a list itself e g FruitFn BananaTree MyBanana this shortcut is not available that is in that case an extra level of list nesting is always necessary The returned truth value represents the logical truth of the queried sentence as determined by PowerLoom It can be be tested via the functions is true is false and is unknown which see Chapter 6 PowerLoom API 41 assert binary proposition relation LOGIC OBJECT arg OBJECT Function value OBJECT module MODULE environment ENVIRONMENT PROPOSITION Assert that the proposition relation arg value is TRUE in module Return the
223. sserted by a defrule is an implication The assertional truth of a proposition defined by a rule can be altered by asserting or retracting the constant that denotes the proposition lt keyword option gt lt keyword gt lt word gt lt definition gt lt reldefinition gt lt objdefinition gt lt ruledefinition gt lt reldefinition gt defconcept lt conceptconst gt lt vardecl gt documentation lt string gt lt lt sentence gt gt lt sentence gt lt lt lt sentence gt gt gt lt sentence gt lt gt lt sentence gt lt gt gt lt sentence gt lt lt gt lt sentence gt lt lt gt gt lt sentence gt axioms lt sentence gt lt sentence gt lt keyword option gt deffunction lt funconst gt lt vardecl gt documentation lt string gt gt lt vardecl gt lt lt sentence gt gt lt sentence gt lt lt lt sentence gt gt gt lt sentence gt lt gt lt sentence gt lt gt gt lt sentence gt lt lt gt lt sentence gt lt lt gt gt lt sentence gt axioms lt sentence gt lt sentence gt lt keyword option gt EENEN C UM CO L3 defrelation lt relconst gt lt vardecl gt documentation lt string gt lt lt sentence gt gt lt sentence gt lt lt lt sentence gt gt gt lt sentence gt
224. ssion 4 2 Persistent Knowledge Bases Serious users of PowerLoom will want to construct knowledge bases that persist between sessions PowerLoom s primary medium of persistence is file based users construct their knowledge bases by entering PowerLoom statements into ASCII formatted files and then using the load command to load them into PowerLoom There is also a save module command that saves the current assertions of a module to a file Large scale persistence via a backend database is currently under development and will become available in one of the next releases Chapter 5 Commands 27 5 Commands This chapter lists all available PowerLoom commands alphabetically Each command is doc umented with its name a possibly empty list of parameters specified as lt name gt lt type gt pairs its return type and its category Command Almost all of the commands implic itly quote their arguments meaning that when calling them you don t need to add any quotation yourself For example the command all facts of is defined as follows all facts of instanceRef NAME CONS OF PROPOSITION Command Return a cons list of all definite TRUE or FALSE propositions that reference the instance instanceRef The all facts of command has one parameter called instanceRef of type NAME and returns a STELLA LIST containing zero or more objects of type PROPOSITION as its result The type NAME subsumes the types SYMBOL SURROGATE STRING
225. state lees 118 Function Index evaluation state setter sss 118 EXAMPLE e1 sm knee Pe Ee Ae REP ei 72 explain why e o e n eee ini 118 explain vhynot i a d di b re ER RES 118 F false truth value s s leeesess 118 false wa iis E E ge erre geen tease 118 fetch instance on OBJECT STORE 118 fetch relation on OBJECT STORE 118 FIFTH ELEMENT os pii eneses e tease idem ae aes eee 72 fill array on 2 D ARRAY sse 118 fill array on 2 D FLOAT ARRAY 118 FILLERS derrer ge achat NP deed arts 12 finalize objects i 0c cae nrs in e eee 118 find direct supers and subs 119 find direct supers of instance 119 find instance veas nerui ey RES 119 find rule ia ne DR PRG cheney Eo ehe 119 EIRSI ELEMENT esheg bib ves Ret er SNES 12 FORK setidir a SEEE EEES Spannung Sete eT REUS 12 FOURTH ELEMENT Rer eenen Reeves T2 FRAME PREDIGATE mei IG PDRere EE IS T2 f nctioh i s e Ra e ERU P erate esa Y 119 G generate expression sses 119 generate specialized term on LOGIC THING iPaaerserhtrigeRkritntkig dad ee er ee ge 119 generate unique name codec eden 43 get arity i e chips estiti deberi E E ER dec 43 get binary proposition sinies 43 get binary propositions sess 43 get child modu les oce ek ena 43 get classS eode ARIA Re bEA E Seeds 119 get column cou
226. static variable Another way to create stella objects is to use edu isi stella Stella unstringify static method This method takes a string representation of a stella object and returns the object If passed a list an object of type Cons will be returned As an alternative one can also convert one and two dimensional arrays of Stella_Object into Cons objects using the overloaded function edu isi stella javalib arrayToCons These functions will return Cons objects constructed from the input arrays Keywords and symbols are objects that are stored in global static variables The variable names are all in upper case and are constructed by concatenating the tag SYM with the module name and the name of the symbol or concatenating the tag KWD with the name of the keyword For example the symbol BACKWARD in the logic module would be stored in edu isi powerloom logic Logic SYM_LOGIC_BACKWARD whereas the keyword ERROR in the stella module would be in edu isi stella Stella KWD_ERROR 6 2 3 5 PowerLoom and Threads The most important consideration when using PowerLoom in a threaded environment is that the core of PowerLoom must not execute in concurrently running threads The PLI class takes care of this for interface functions that run through that class Other PowerLoom functions that are called need to synchronize on a lock object edu isi powerloom logic Logic POWERLOOM LOCK for proper operation This is not needed for setting Special Variables
227. stream OUTPUT STREAM Method on NAMED DESCRIPTION mode KEYWORD Prints a description of self to stream stream mode can be terse verbose or source Used by describe description name self SYMBOL Method on NAMED DESCRIPTION Return the name of the description self Chapter 11 Miscellaneous 117 description name self SYMBOL Method on DESCRIPTION Return the name of the description self if it has one destroy instance self OBJECT Function Destroy all propositions that reference self and mark it as deleted thereby making it invisible within class extensions destroy object self OBJECT Function Destroy self which can be a term or a proposition Destroy all propositions that reference self and mark it as deleted thereby making it invisible within class extensions destroy proposition proposition PROPOSITION PROPOSITION Function Retract and destroy the proposition proposition Recursively destroy all propositions that reference proposition Also destroy all satellite propositions of proposition destroy term self LOGIC OBJECT Function Destroy all propositions that reference self and mark it as deleted thereby making it invisible within class extensions Unlink descriptions from native relations direct superrelations self RELATION ITERATOR OF LIKE SELF Function Return direct super classes slots of self disabled powerloom feature feature KEYWORD BOOLEAN Function
228. t exception is thrown s get relation name STRING module name STRING Function environment ENVIRONMENT LOGIC OBJECT Return a concept or relation named name that is local to or visible from the module module name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown Chapter 6 PowerLoom API 57 s get relation extension relation name STRING module MODULE Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return propositions that satisfy the relation named relation name Include proposi tions that satisfy subrelations of the relation s get rules relation name STRING module name STRING Function environment ENVIRONMENT PL ITERATOR OF PROPOSITION Return rules attached to the concept relation named relation name found in the module named module name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella No Such Context Exception is thrown s is true proposition relation and arguments STRING Function module name STRING environment ENVIRONMENT BOOLEAN Return TRUE if a proposition that prints as the string relation and arguments is true in the module named module name A module name of null or the empty string refers to the current module If no module can be found with the name modu
229. t Mary is a person the answer will be yes TRUE because Mary is asserted to be an employee and membership in employee implies membership in person In other words although the direct assertion that Mary is a person is not present in the knowledge base a logical proof exists that the proposition Mary is a person is true 2 4 Modules The knowledge loaded into an executing PowerLoom system is divided into logical partitions called modules The modules are arranged into a hierarchy knowledge inherits down the hierarchy from parents to children A convenient way to organize knowledge is to put definitional knowledge higher up in the module hierarchy and factual knowledge lower down For example suppose we want to build a knowledge base that defines a business domain and include a substantial number of facts about individual companies We might use one or a few modules to define terminology that relates to the business domain and then places the set of facts about each company in its own module If we were querying the knowledge base about one or a few companies it would not be necessary to load the modules for the remaining companies into the system Facts asserted within a module are not visible in sibling modules or in ancestor modules Thus if we enter into PowerLoom an assertion that Georgia is a state we are not asserting that Georgia is a state in all possible worlds but that from the vantage point of the current module
230. t function slot relation SURROGATE BOOLEAN Function Return TRUE if relation a surrogate is a function test marker self object OBJECT BOOLEAN Method on MARKER TABLE Return TRUE if object is stored marked in self test relation on arguments relation SURROGATE Function arguments CONS BOOLEAN Return TRUE if relation a surrogate is TRUE when applied to arguments Chapter 11 Miscellaneous 129 test slot value self LOGIC OBJECT relation SURROGATE Function filler OBJECT BOOLEAN Return TRUE if the proposition lt relation gt lt self gt lt filler gt is true test special marker table self OBJECT BOOLEAN Function Return TRUE if the object self is stored marked in the table pointed at by the special variable specialMarkerTable Designed for use by remove if test subrelation subrelation SURROGATE Function superrelation SURROGATE BOOLEAN Return TRUE if subrelation specializes superrelation test type on instance self OBJECT type SURROGATE BOOLEAN Function Return TRUE if self satisfies type translate loom file input FILE NAME output FILE NAME Command Translate the Loom file input to PowerLoom and write the translation to the file output Note that this will only work for fairly vanilla Loom files that do not contain any Lisp isms It might require to clean the Loom file manually before this translation will work true truth value self TRUTH VALU
231. t optimize true false confidence level strict default lt keyword option gt lt sentence gt can be any sentence that is legal as a top level assertion lt name gt can be a string or symbol and will be bound to the asserted proposition represented by lt sentence gt After this definition every occurrence of lt name gt will be replaced by the associated proposition The options forward only and backward only can be used to tell the inference engine to only use the rule in forward or backward direction this can also be achieved by using the lt lt or gt gt implication arrows dont optimize tells the inference engine to not rearrange the order of clauses in the antecedent of a rule and instead evaluate them in their original order confidence level can be used to mark a proposition as default only lt keyword option gt represents a keyword followed by a value that states an assertion about the proposition lt name gt See defrelation for a description of lt keyword option gt s defrelation amp rest args PARSE TREE NAMED DESCRIPTION N Command Define or redefine a logic relation The accepted syntax is defrelation lt relconst gt lt vardecl gt documentation lt string gt lt lt sentence gt gt lt sentence gt lt lt lt sentence gt gt gt lt sentence gt lt gt lt sentence gt lt gt gt lt sentence gt lt lt gt lt sen
232. tamorphic Rock and Sedimentary Rock together form a disjoint covering of the concept Rock DOCUMENTATION x THING s STRING Relation s is a documentation string attached to x Some of the PowerLoom text processing tools look for documentation strings and import them into documents DOMAIN r RELATION d COLLECTION Relation True if for any tuple T that satifies r the first argument of T necessarily belongs to the concept d domain exists for convenience only and is defined in terms of nth domain domain assertions should be avoided since they create redundant nth domain propositions use nth domain directly Chapter 7 Built In Relations 72 DUPLICATE FREE c COLLECTION Relation 7c is duplicate free if no two members denote the same object DUPLICATE FREE COLLECTION c COLLECTION Concept 7c is free of duplicates EMPTY c COLLECTION Relation The collection c is empty if it has no members Note that for collections possessing open world semantics e g most concepts the fact that the collection has no known members does not necessarily imply that it is empty EQUIVALENT RELATION r RELATION equiv RELATION Relation True if r is equivalent to equiv written in set notation r equiv This relation will generate bindings for at most one unbound argument EXAMPLE r RELATION e THING Relation 7e is an example of the use of r FIFTH ELEMENT l COLLECTION gt e THIN
233. tation a function that evaluates an atomic relation proposition during query processing The function is passed a proposition for evaluation for which all arguments are bound The function returns a BOOLEAN if it represents a predicate or some sort of value if it is a function RELATION CONSTRAINT r RELATION Relation computation COMPUTED PROCEDURE Names a computation a function that evaluates an atomic relation proposition during query processing The function is passed a proposition for evaluation for which Chapter 7 Built In Relations 80 at most one argument is unbound The function returns a BOOLEAN if it represents a predicate or some sort of value if it is a function If all arguments are bound the function computes whether the constraint holds If all but one argument is bound and the unbound argument is a pattern variable then the missing value is computed RELATION EVALUATOR r RELATION 7ev COMPUTED PROCEDURE Relation Names an evaluator a function that evaluates an atomic relation proposition during constraint propagation This defines an extensible means for computing using auxiliary data structures The function is passed a proposition for evaluation which might update the proposition generate additional assertions or trigger further eval uations Evaluators have to check the truth value of the passed in proposition and perform their actions accordingly An evaluated proposition might be true false or
234. ted a default preferences file is Chapter 8 PowerLoom GUI 90 used The preferences file includes among other things the PowerLoom server that was last accessed 4 If the preferences file contains the last accessed server the GUI attempts to connect to the server and query the server for a description of the server s capabilities If connection is successful a browser window will open displaying the modules that are currently loaded in the server instance 5 The user selects any KB files s he wishes to load and instructs the server to load the files 6 The user performs some browsing querying and editing of the loaded KB 7 If any changes were made the user saves the KB to a new or existing file 8 The user repeats steps 5 7 as needed and then exits the application 8 4 GUI Features This section provides a detailed description of the features that are available in the GUI application We describe general application wide functionality as well as the functionality of specific components 8 4 1 Connect to Server The first time the GUI is started it will not attempt to connect to any server To establish a server connection the user must select the KB Connect to Server menu item This will open a dialog prompting for a host name and port After the user enters this information a connection will be attempted If the connection is successful the server information will be stored in the preferences file and used next
235. tence gt lt lt gt gt lt sentence gt axioms lt sentence gt lt sentence gt SS A Chapter 5 Commands 34 lt keyword option gt Relation parameters can be typed or untyped lt keyword option gt represents a key word followed by a value that states an assertion about lt relconst gt For example including the option foo bar states that the proposition foo lt relconst gt bar is true foo bar fum states that both foo lt relconst gt bar and foo lt relconst gt fum are true foo true states that foo lt relconst gt is true foo false states that not foo lt relconst gt is true defrule amp rest args PARSE TREE PROPOSITION N Command Define or redefine a named rule defrule is an alias for defproposition which see delete rules relation NAME N Command Delete the list of rules associated with relation This function is included mainly for debugging purposes when a user wants to verify the behavior of different sets of rules demo amp rest fileandpause OBJECT Command Read logic commands from a file echo them verbatimly to standard output and evaluate them just as if they had been typed in interactively When called with no arguments present a menu of example demos otherwise use the first argument as the name of the file to demo Pause for user confirmation after each expression has been read but before it is evaluated Pausing can be turned off by supp
236. ter A string lt string gt is a character sequence including words plus all special charcters except double quotes enclosed in double quotes A double quote can be included in a string if it is preceeded by the escape character 13 2 Grammar Legal expressions in PowerLoom are forms which are either a statement or a definition described in more detail below lt form gt statement definition 13 2 1 Constants and Typed Variables The language consists of several groups of operators defined as follows lt termop gt listof setof the setofall kappa lt sentop gt not and or forall exists lt E gt lt lt 325 gt lt lt 7 lt defop gt defconcept deffunction defrelation defrule documentation gt lt gt lt lt gt gt lt gt lt gt gt lt lt gt lt lt gt gt axioms lt operator gt termop lt sentop gt lt defop gt All other words are constants words which are not operators or variables constant word lt indvar gt lt operator gt Semantically there are different categories of constants Concept constants lt conceptconst gt Function constants lt funconst gt Relation constants relconst Rule Chapter 13 PowerLoom Grammar 138 constants lt ruleconst gt and Logical constants lt logconst gt The differences between these categories are
237. ters and return a boolean The code will use the appropriate test for the specific subtype of wrapped number actually passed in define computed constraint name SYMBOL var list CONS Macro constraint test CONS amp body position computations CONS OBJECT Defines name to be a constraint computation which uses constraint test to deter mine if a fully bound set of variables satisfies the constraint The forms in position computations are used to compute the value for each of the positions All such computations must set the variable value to be the result computed for the missing position Setting value to nu11 for any such computation means that that particular argument cannot be computed from the others The input variables in var list will be bound to the N arguments to the constraint The generated function will return a Stella Object and take as inputs the values of the N arguments to the constraint A value of null means that the value is not available If all arguments are not null then the return value will be a Stella wrapped boolean indicating whether the con straint is satisified or not If more than one input value is nu11 then this constraint code will not be called deobjectify tree self OBJECT OBJECT Function Return a copy of self where all logic objects are replaced by their generated parse tree version This is useful to convert the result of a retrieval query into a regular parse tree describe object self
238. tion is false with full inference effort ask its negation KIF example ask happy Fred will return TRUE if Fred was indeed found to be happy Note that for this query to run the logic constant Fred and the relation happy must already be defined see assert Use set unset feature goal trace to en disable goal tracing of the inference engine The ask command supports the following options TIMEOUT is an integer or floating point time limit specified in seconds For example the command ask nervous Fred timeout 2 0 will cease inference after two seconds if a proof has not been found by then If the DONT OPTIMIZE is given as TRUE it tells PowerLoom to not optimize the order of clauses in the query before evaluating it This is useful for cases where a specific evaluation order of the clauses is required or the optimizer doesn t do the right thing If THREE VALUED is given as TRUE PowerLoom will try to prove the negation of the query with full effort in case the given query returned UNKNOWN By default PowerLoom uses full effort to prove the query as stated and only a little opportunistic effort to see whether it is actually false assert proposition PARSE TREE OBJECT N Command Assert the truth of proposition Return the asserted proposition object KIF example assert happy Fred asserts that Fred is indeed happy Note that for this assertion to succeed the relation happy must already be defined If the constant F
239. tion specialist lookup native specialist native name STRING FUNCTION CODE Function Returns the native funtion code for native name if it exists and the underlying pro gramming languages supports such lookups Uses the signature of a specialist func tion Chapter 11 Miscellaneous 123 named description self DESCRIPTION BOOLEAN Function Return TRUE if self is the description of a named class or relation natural deduction mode BOOLEAN Function True if normalization is governed by natural deduction semantics non empty self BOOLEAN Method on QUERY SOLUTION TABLE Return TRUE if self has at least 1 entry non empty self BOOLEAN Method on FLOAT VECTOR Return TRUE if self has length gt 0 nth self position INTEGER LIKE Method on QUERY SOLUTION TABLE ANY VALUE SELF Return the nth solution in self or NULL if it is empty object name self OBJECT SYMBOL Function Return the name symbol for the logic object self object name string self OBJECT STRING Function Return the name string for the logic object self object surrogate self OBJECT SURROGATE Function Return the surrogate naming the object self which may be a Stella class that is used in PowerLoom as well as a more normal powerloom object object surrogate setter self OBJECT name SURROGATE Function SURROGATE Return the name of the logic object self to name pop self LIKE ANY VALUE SELF Met
240. to that module unless an in module declaration is encountered which will over ride the default value If noe default module is specified and the input stream does not have an in module form an error is signaled logic class self CLASS BOOLEAN Function Return TRUE if the class self or one of its supers supports indices that record ex tensions referenced by the logic system Also return true for literal classes logic form less ol OBJECT 02 OBJECT BOOLEAN Function A sorting predicate for objects o1 and o2 that can appear in logical forms Performs a combined numeric and lexocographic sort that accounts for lists collections and propositions Numbers precede all other values nu11 follows all other values logic module self MODULE BOOLEAN Function Return TRUE if self is a logic module implying that relations defined within it define a knowledge base A module is a logic module iff it inherits the module PL KERNEL lookup self key LIKE ANY KEY SELF Method on QUERY SOLUTION TABLE LIKE ANY VALUE SELF Lookup the solution identified by key in self and return its value or NULL if no such solution exists lookup native computation native name STRING arity INTEGER Function FUNCTION CODE Returns the native funtion code for native name if it exists and the underlying pro gramming languages supports such lookups It is looked up using the signature of a computation function supported by the computa
241. to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get concept instances concept name STRING Function module name STRING environment ENVIRONMENT PL ITERATOR Return instances of concept concept name Include instances of subconcepts of concept name Depending on concept name the return values could be wrapped literals A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get direct concept instances concept name STRING Function module name STRING environment ENVIRONMENT PL ITERATOR Return instances of concept concept name Exclude instances of subconcepts of concept name Depending on concept name the return values could be wrapped literals Chapter 6 PowerLoom API 55 A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is thrown s get domain relation name STRING module name STRING Function environment ENVIRONMENT LOGIC OBJECT Return the type concept for the first argument to the binary relation relation name A module name of null or the empty string refers to the current module If no module can be found with the name module name then a Stella no such context exception is throw
242. try is that checking for truth and falsity really amounts to asking two separate and possibly expensive queries and the user or programmer should decide whether the effort should be expended to ask both queries instead of just one PowerLoom can sometimes infer a negative fact without the necessity of a direct asser tion For example ask number of employees ACME cleaners 8 gt TRUE ask number of employees ACME cleaners 10 gt FALSE ask not number of employees ACME cleaners 10 TRUE PowerLoom can infer the second and third answers because it knows that the function number of employees can return only one value and if that value is the number eight it cannot also be something else in this case ten Many systems in particular database systems and Prolog make the assumptions that if a proposition cannot be proved true then it must be false This is called the closed world assumption By default PowerLoom makes an open world assumption but for specific relations it can be instructed to assume a closed world if a user wants closed world semantics For example suppose we introduce a relation works for and we assume that all works for facts have been entered in our knowledge base defrelation works for p c Company assert works for shirly ACME cleaners assert works for jerome zz productions If we ask PowerLoom whether Jerome does NOT work for MegaSoft it will return unkn
243. ts a b and c on successive iterations MAXIMUM ELEMENT c COLLECTION sortby RELATION Relation element THING Find those element s of c with the maximum sortBy value for example retrieve maximum element relation arity x see also collect into descending set MAXIMUM VALUE COLLECTION max NUMBER Relation Binds max to the maximum of the numbers in the list 1 MEAN VALUE COLLECTION mean NUMBER Relation Binds mean to the mean of the numbers in 71 MEDIAN VALUE COLLECTION median NUMBER Relation Binds median to the median of the numbers in 71 MEMBER OF x THING c COLLECTION Relation TRUE if x is a member of collection c common use of member of is for binding a variable to successive members in a list or set see listof and setof Chapter 7 Built In Relations 76 MINIMUM ELEMENT c COLLECTION sortby RELATION Relation element THING Find those element s of c with the minimum sortBy value for example retrieve minimum element relation arity x see also collect into ascending set MINIMUM VALUE l COLLECTION min NUMBER Relation Binds min to the minimum of the numbers in the list MUTUALLY DISJOINT COLLECTION s sET Relation True if the members of s are pair wise disjoint Used most often to expresse disjoint ness constraints between concepts For example mutually disjoint collection setof MAN WOMAN states that th
244. ude instances of subconcepts of concept Depending on concept the return values could be wrapped literals get direct subrelations relation LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF LOGIC OBJECT Return relations that directly specialize relation Non reflexive get direct superrelations relation LOGIC OBJECT Function module MODULE environment ENVIRONMENT PL ITERATOR OF LOGIC OBJECT Return relations that directly generalize relation Non reflexive get direct types object LOGIC OBJECT module MODULE Function environment ENVIRONMENT PL ITERATOR OF LOGIC OBJECT Return most specific concepts that object belongs to get domain relation LOGIC OBJECT LOGIC OBJECT Function Return the type a concept for the first argument to the binary relation relation get enumerated collection members collection OBJECT Function module MODULE environment ENVIRONMENT CONS Returns the members of an enumerated collection This works on all types of collec tion i e sets and lists Chapter 6 PowerLoom API 45 get home module object LOGIC OBJECT MODULE Function Return the module in which object was created get inferred binary proposition values Function relation LOGIC OBJECT arg OBJECT module MODULE environment ENVIRONMENT PL ITERATOR Return all values v such that relation arg v has been asserted or can be inferred get module name STRI
245. ule name NAME MODULE N Command Change the current module to the module named name list modules kb only BOOLEAN CONS OF MODULE Command Returns a cons of all modules defined in PowerLoom If kb only is true then any modules which are code only or just namespaces are not returned load file STRING amp rest options OBJECT Command Read logic commands from file and evaluate them By default this will check for each asserted proposition whether an equivalent proposition already exists and if so not assert the duplicate These duplicate checks are somewhat expensive though and can be skipped by setting the option check duplicates to false This can save time when loading large KBs where it is known that no duplicate assertions exist in a file Also by setting the option module the module in which the file contents will be loaded will be set This will only affect files that do NOT have an in module dec laration as part of the file If this is not set and no in module declaration is in the file then an error will be signaled load file file STRING Command Read STELLA commands from file and evaluate them The file should begin with an in module declaration that specifies the module within which all remaining com mands are to be evaluated The remaining commands are evaluated one by one ap plying the function evaluate to each of them pop load path STRING Command Remove the first element from the
246. un forward rules temporarily puts the module into a mode where future assertional monotonic updates will trigger additional forward inference Once a non monotonic update is performed i e a retraction or clipping of relation value all cached forward inferences will be discarded and forward inferencing will be disabled until this function is called again run powerloom tests Command Run the PowerLoom test suite Currently this simply runs all demos and echos commands and their results to standard output The output can then be diffed with previously validated runs to find deviations satisfies instanceOrTuple OBJECT relationRef OBJECT Function TRUTH VALUE Try to prove whether instanceOrTuple satisfies the definition of the relation relation Ref and return the result truth value of the query instanceOrTuple can be a single object the name or surrogate of an object or a collection a list or vector of objects relationRef can be a relation description surrogate or relation name save all neural networks file STRING Command Save all neural networks to file if file is non NULL If networks are saved periodically see set save network cycle this file name will be used to perform periodic saves Chapter 11 Miscellaneous 127 select proof result success BOOLEAN continuing BOOLEAN Function terminal BOOLEAN KEYWORD Helping function for specialists Return the appropriate keyword indicating success
247. uring that users don t accidentally step on each other We need to improve our infrastructure to handle Chapter 8 PowerLoom GUI 102 notification of KB updates add support for transactions and KB locking and improve our caching mechanisms to detect when the GUI state is out of sync with respect to the server Chapter 9 Ontosaurus 103 9 Ontosaurus Ontosaurus is a Web based KB browser that offers a convenient way to view the contents of PowerLoom knowledge bases from a standard Web browser Whenever a user searches for an object or clicks on a hyperlink Ontosaurus dynamically generates HTML pages describing the requested object and related information The description will itself contain links to other relevant objects which the user can click on to quickly surf through a knowledge base Similar to the PowerLoom GUI see Chapter 8 PowerLoom GUI page 84 Ontosaurus relies on a PowerLoom server that can handle HTTP requests For this reason it is currently only supported in the Java version of PowerLoom or in Lisp version that supports a Web server such as AServe or Portable AServe e g Allegro CL To use Ontosaurus do the following Start PowerLoom in Java or a qualifying Lisp version Once PowerLoom is up and running issue the following command PL USER start ontosaurus Loading required system webtools Loading required system ontosaurus PL USER The command also accepts a port option to make the server listen to a
248. us scraps of information We envision a scrapbook feature where text and objects of various types could be dragged to and organized visually for later use 8 5 5 Instance Cloning It is often useful to create new instances that are similar to existing instances We would like to implement a cloning facility in which a wizard like series of dialogs would step the user through the process of copying information from one object to a new object For example the dialogs would prompt the user for propositions to transfer from the old instance to the new instance and allow the user to modify the propositions in the process of transferring them 8 5 6 Security There is currently no authentication and access control mechanisms in the PowerLoom GUI The GUI client assumes that is communicating with a trusted host over a secure network Similarly the PowerLoom server assumes that it is communicating with a friendly client that has full access to the server In the future we need to add security mechanisms which allow clients to be authenticated and resources on the server to be made accessible to authorized users only In addition we need to implement encryption mechanisms so that clear text is not sent over insecure networks potentially compromising sensitive data 8 5 7 Multiple Users Although the client server model allows multiple GUI clients to concurrently share the same server there is only very weak support for synchronizing clients and ens
249. uses mixed or lower case The mappings use lowercase characters if the basic identifier uses upper case Stella Java Mnemonic gt X eXclamation ES _ gt 4H Hash gt B Buck gt R peRcent amp gt A Ampersand gt 0 Quote i a L gt gt I Increase gt _ gt _ gt D Dot S Slash gt C Colon T gt uu lt L Less than E Equal gt gt G Greater than gt P Predicate gt M Monkey tail J Arbitrary array index ME M gt K Arbitrary array index U Up arrow c gt _ Y Arbitrary adjacent free letter gt V Vertical bar Z Arbitrary adjacent free letter m Tilde Chapter 6 PowerLoom API 66 lt space gt gt _ 6 2 3 8 Stella Exceptions in Java Stella exceptions are implemented as a subtype of java lang Exception actually RunTimeException and may be caught normally All Stella Exceptions belong to the edu isi stella StellaException class or one of its subclasses The more specfic PowerLoom exceptions belong to the edu isi powerloom logic LogicException class or one of its subclasses Exceptions thrown during I O operations will not use the standard Java excep tions Instead they will descend from edu isi stella Input utputException The most useful descendents are edu isi stella NoSuchFileException and edu isi stella EndOfFileException 6 2 3 9 Iteration in Java Iteration in Ste
250. valuated or NULL if it has never been evaluated evaluation state setter proposition PROPOSITION Function state KEYWORD Record the evaluation state of proposition explain why label STRING style KEYWORD maxdepth INTEGER Function stream OUTPUT STREAM Programmer s interface to WHY function explain whynot label STRING style KEYWORD maxdepth INTEGER Function summary BOOLEAN stream OUTPUT STREAM Programmer s interface to the WHYNOT function false truth value self TRUTH VALUE BOOLEAN Function Return TRUE if self represents some form of falsehood false self PROPOSITION BOOLEAN Function Return true if self is false or default false if we are considering default assertions fetch instance store name OBJECT OBJECT Method on OBJECT STORE Fetch the instance identified by name a string or symbol from store and return it as an appropriate logic object This needs to be appropriately specialized on actual OBJECT STORE implementations fetch relation store name OBJECT Method on OBJECT STORE NAMED DESCRIPTION Fetch the relation identified by name a string or symbol from store and return it as a named description This needs to be appropriately specialized on actual OBJECT STORE implementations fill array self amp rest values OBJECT Method on 2_D_ARRAY Fill the two dimensional array self in row major order from values Missing values will retain their o
251. value Note that similar to other collect into functions c can be a named concept a setofall or an enumer ated collection and sortBy can be a named relation or a kappa COLLECT INTO LIST c COLLECTION gt 71 LIST Function Infer as many members of c as possible and collect them into a list 1 This is similar to collect into set but collects members in the order they are encountered If a Chapter 7 Built In Relations 70 member is derived multiple times all occurrences are kept This is useful for example to collect and then sum up number valued attributes of objects COLLECT INTO ORDERED SET c COLLECTION gt LIST Function This is similar to collect into list but if a member is derived multiple times only the first occurrence is kept hence the name even though ordered sets are represented as lists COLLECT INTO SET c COLLECTION gt SET Function Infer as many members of c as possible and collect them into a set l For example here is a rule used to compute bindings for the fillers predicate lt fillers r i members and bound variables r i collect into set setofall v holds r i v members When r and i are bound the term setofall v holds r i v evaluates to a unary relation satisfied for each filler of the relation in r applied to the instance in i collect into set causes the extension of this dynamically defined unary relation to be computed
252. ve list of procedures that can be called to control the logic system These procedures range from very specific procedures to assert or query a single fact to general procedures that interpret arbitrary queries The STELLA translator offers users a choice of Common Lisp Java or C based versions of the PowerLoom system users can choose whichever is the best match for their language of choice for their applications Starting with PowerLoom version 4 0 an experimental Java based GUI is available to browse query and edit knowledge bases see Chapter 8 PowerLoom GUI page 84 Ad ditionally the Ontosaurus Web Browser offers a convenient way to view the contents of PowerLoom knowledge bases from a standard Web browser see Chapter 9 Ontosaurus page 103 4 1 Command Interpreter Currently the primary means for interacting with PowerLoom is its command interpreter The command interpreter can be used either interactively or it can be invoked directly from a program to evaluate individual commands All PowerLoom commands see Chapter 5 Commands page 27 can be evaluated using the command interpreter The interactive command interpreter is invoked by calling the function powerloom with out any arguments In the Java versions of PowerLoom the interpreter in called by the main routine in the class PowerLoom within the logic package In the C versions of PowerLoom powerloom is also called within the main routine In the Lisp version S
253. will return the pre existing proposition Multiple propositions may be returned for a single sentence because of normalization of equivalences conjunctions etc Signals a Proposition Error if PowerLoom could not conceive sentence cons to pl iterator self CONS PL ITERATOR Function Convert a Stella cons list into an API iterator create concept name STRING parent LOGIC OBJECT Function module MODULE environment ENVIRONMENT LOGIC OBJECT Create a concept named name in the designated module with the designated parent superconcept which can be left undefined Additional superconcepts can be added via assertions of the subset of relation Note that a specified parent concept needs to be created separately Chapter 6 PowerLoom API 42 Note that because names in modules that are not case sensitive are canonicalized the name of the returned object may not match name exactly create enumerated list members CONS module MODULE Function environment ENVIRONMENT LOGIC OBJECT Create a logical term that denotes a list containing members in module using envi ronment Useful for passing lists as arguments to parameterized queries create enumerated set members CONS module MODULE Function environment ENVIRONMENT LOGIC OBJECT Create a logical term that denotes the enumerated set containing members in module using environment create function name STRING arity INTEGER module MODULE Function environ
254. y of relations A proposition is a logical sentence that has an associated truth value Examples are Ben Franklin is a person Bill is married to Hillary Two plus three equals six which is false PowerLoom follows KIF in adopting a prefix notation for the predicate calculus to represent propositions Possible representations of the three propositions just mentioned are person ben franklin married to Bill Hillary and 2 3 6 These three propositions make reference to relations named person married to plus and The predicate calculus constructs complex sentences out of simpler ones using the log ical connectives and or not lt gt and lt gt and the quantifiers exists and forall Some examples are not crook richard Richard is not a crook and forall p gt person p exists m has mother p m every person has a mother Chapter 2 Conceptual Framework 7 2 2 Definitions PowerLoom requires that relations are defined before they are used within assertions and queries The commands defconcept defrelation and deffunction are used to define concepts relations and functions respectively The definitions defconcept person defrelation married to p1 person p2 person deffunction n1 number n2 number gt sum number declare that person is a concept that married to is a binary relation that takes ar guments of type person and that is a functi

from isi.edu - Information Sciences Institute

Contents

Download Pdf Manuals

Related Search

Related Contents