Home

The Elements of AIML Style

1. 45 Th systerms COPImqntl eoque Qupa quni 45 THE SHBEDLSQREE co ca bees be itas as s aset ias 46 THE PROLOG PROGRAM aa a Dan naui dosi RU d 46 THE AME CA TBGOBRIBS oi eU aprendo 48 SAMPLE 0017 2552 59222228 cma ed A u AL e 48 ACTIVATING PROLOG FROM tapes te eoi ted ases dit Mee Period eR Seed 49 PROLOG ASSERTIONS eet aperte 50 CHAPTER VI Doing Simple Logical Deductions in AIML seen 52 te n ie ue Seem Ce t 52 KNOWLBDGHB us Sui 53 DEBAULETS a pd ELE E MM 54 REDUCTIONS eti d tatu dde ado tatu 54 DEDUCHONS n m Ra 55 55 RANDOM INFERENCES 56 SAMPLE DIALOG tdt mi Huet sedan 57 LOGICAL DEDUCTIONS 58 CHAPTER VII A Personal Finance Spreadsheet in AIML 58 SAMPLE DIALOG qst o 58 THE TRANSACTIONS LAS A Qum sis
2. 21 THE ALICE BRAIN PICTURKEGALPEEBRY aude 23 CHAPTER III Symbolic Reductions in 4 8000 33 TIPE WILDDCARDS de E Ie IN mu de 35 CHAPTER IV Botmaster Questions and Answers eene ener nne 36 lt What 1s the Tor 36 WG S the a s m 37 How can I create my own chat robot ee ae ee Bs 37 How difficult is it to create a chat robot 37 lt Does ALICE learn iy ut nn SS su t oti mie uui une 38 Does ALICEK a e S kaun a 38 What IS the theory behind ALICE u 38 Why Is AIML Category Called A Category 39 Can I have a private conversation with ALICE nennen 40 How do I catch all occurrences of a keyword in the input 40 s How Abe Lar els generated uote ul a en M d aUe anc buco doux 41 SOME EXAMPLES a uu s 42 CHAPTER V PSYCH Activating Prolog from AIML
3. BOUGHT FOR DOLLARS ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star index 2 gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt lt aiml gt CHAPTER VIII Building a Subscriber Based Bot Business with Pandorabots If was making the film The Graduate with Dustin Hoffmann today instead of advising him to go into plastics would say bots would say forget about graduate school and academia that is a dead end Go to work for yourself if you can and join the A L I C E and AIML free software community You will have the benefit of a huge free software development effort and community free and be able to devote most of your efforts to marketing promotions advertising and sales Seriously if was a young healthy person like yourself would go for it too Here is a list of the top 10 killer apps for AIML put together Ten Killer Apps for A L 1 C E and AIML 1 ESL Bot ESL sites are the 1 source of traffic to Alicebot org ESL students need a lot of practice chatting in English and the bot provides a very safe environment to practice 2 FAQbot Any FAQ can be converted to AIML easily Any organization or individual answering a set of frequently asked questions can reduce their workload with a bot 3 Toy Combined with speech recognition creates talking teddy bea
4. 40 How are targets generated The perfect targeting algorithm has not yet been developed Meanwhile we rely on heuristics to select targets from the activated categories The A L C E brain at the time of this writing contains around 40 000 categories In any given run of the server however typically only a few thousand of those categories are activated Potentially every activated category is a source of targets If more than one input activated some category then each of those inputs potentially forms a new target The first step in targeting is to save all the activated categories and the inputs that activated them Program dB is an experimental Alicebot engine and server that includes code from both Program D and Program B This is the source code of the program that running at the Loebner contest held at the London Science Museum on 13 October 2001 where A L I C E won her second Loebner prize Program B was the original Java version of the Alicebot engine and server When Jon Baer took over Java development 2001 and subsequently the A L I C E Al Foundation sponsored the development of the AIML 1 0 reference server these efforts came to be known as Program D Program C refers to the collection of AIML programs written in C C Program dB has its own built in automatic targeting algorithm Program dB applies six steps to each activated category to either accept it as a source of targets
5. subfolder Forgetting about that and topic for a minute we can imagine that the folder g I LIKE TO star has a single file called template txt that contains the template If all the patterns and templates are placed into the file system in that way we can easily rewrite the explanation of the matching algorithm If you are given an input starting with word X and a folder of the filesystem If the input is null and the folder contains the file template txt halt Does the folder contain the subfolder underscore If so change directory to the underscore subfolder Try all remaining suffixes of the input following X to see if one matches If no match was found try Does the folder contain the subfolder X If so change directory to the subfolder X using the tail of the input the suffix of the input with X removed If no match was found try Does the folder contain the subfolder star If so change directory to the star subfolder Try all remaining suffixes of the input following X to see if one matches If no match was found change directory back to the parent of this folder and put X back on the head of the input Editor s note underscore and star as directory names above are meant to stand in for and which are not allowed as file or directory names in some operating Systems Since the literals underscore and star might be actual words in a pattern perhaps a real implement
6. What else do X have If the program chooses the second form the question itself becomes the input to this same category category pattern WHAT ELSE DO DO lt pattern gt template random li srai WHAT DO lt star gt DO lt srai gt lt li gt li srai WHAT ELSE DO srai ISB lt star gt lt srai gt DO lt srai gt lt li gt lt random gt lt template gt lt category gt lt category gt lt pattern gt WHAT ELSE HAVE lt pattern gt lt template gt lt random gt li srai WHAT DO lt star gt HAVE lt srai gt lt li gt lt li gt lt srai gt WHAT ELSE DO lt srai gt ISB star srai HAVE lt srai gt lt li gt lt random gt lt template gt lt category gt SAMPLE DIALOG The following dialog demonstrates the categories described in this technical note Client What does a bird have Robot Feathers Client What does a bird have Robot Lungs Client What does a bird have Robot A beak Client What do birds have Robot A beak Client What does a raven do Robot Fly Client What else does a bird have Robot A tail Client What else does a bird have Robot Nuclear membranes Client What else does a bird have Robot A body Client What else does a bird have Robot Imagine no possessions Client What else does a chicken do Robot Get eaten Client What else does a chicken do Robot Cluck Client What else do chicken
7. One ambitious goal for AIML is that if a number of people create their own robots each with a unique area of expertise program B can literally merge sort them together into a Superbot automatically omitting duplicate categories We offer the both the source code and the ALICE 36 content in order to encourage others will open source their chat robots as well to contribute to the Superbot Botmasters are also of course free to copy protect private chat robots Who is the botmaster The botmaster is you the master of your chat robot A botmaster runs program B and creates or modifies a chat robot with the program s graphical user interface GUI He or she is responsible for reading the dialogues analyzing the responses and creating new replies for the patterns detected by the AIML server Botmasters are hobbyists webmasters developers advertisers artists publishers editors engineers and anyone else interested in creating a personal chat robot How can create my own chat robot The secret to chat bot programming if there is one is what Simon Laven called continuous beta testing Program B runs as a server and collects dialog on the web The program provides the chat bot developer with a tool called classify dialogues that tests the current robot with the history of accumulated human queries Moreover the program suggests new categories automatically for the botmaster to refine How difficult is i
8. R Ha ha very funny Nancy Internally the AIML interpreter stores the input pattern that pattern and topic pattern along a single path like INPUT that THAT topic TOPIC When the values of that or topic are not specified the program implicitly sets the values of the corresponding THAT or TOPIC pattern to the wildcard The first part of the path to match is the input If more than one category have the same input pattern the program may distinguish between them depending on the value of that If two or more categories have the same pattern and that the final step is to choose the reply based on the topic This structure suggests a design rule never use that unless you have written two categories with the same pattern and never use topic unless you write two categories with the same pattern that Still one of the most useful applications for topic is to create subject dependent pickup lines like topic name CARS gt category lt pattern gt lt pattern gt lt template gt lt random gt lt li gt What s your favorite car lt li gt lt li gt What kind of car do you drive lt li gt lt li gt Do you get a lot of parking tickets lt li gt lt li gt My favorite car is one with a driver lt li gt lt random gt lt template gt Considering the vast size of the set of things people could say that are grammatically correct or semanticall
9. The Elements of AIML Style March 28 2003 By Dr Richard S WALLACE 2003 ALICE A I Foundation Inc The Elements of AIML Style is a no nonsense technical book that takes you on a journey from the first steps of creating your own bot with AIML through all the questions and answers every botmaster asks to advanced and hard nosed business applications of AIML The trip ends with a brief glimpse into the future of bots and AIML WHADPMAKES ALICE WORK etti up pite up E 6 AREVOLUTION IN ND O Su tutta este M uu 10 CHAPTER L Overview 0f Opes ea ute iE eee 12 M disc dir n 12 RECURSIQN osa souci tee ed been ha uk de e hay Chua e Red 13 SYMBOLIC REDUCTIONS ORE HER TO MOM EUR 13 DIVIDE AND GONQUER 5 e ang ei ue dist Ud den bad i iR ue 14 SYNONYMS gunu qaqa epp aU EDO naha naa 14 SPELLING AND GRAMMER CORRECTION 15 KEYWORDS 2260 e a AN RUE Ud tu area e EI UR 15 CONDITIONES ODD RE DR RUD AU Rr MEN LM OR 16 TARGETING nonoina eren t D Pu d UR HE OUS DUREE 16 CON EB 17 CHAPTER Pattern Matching asas ceti uen ed 21
10. gt PAID FOR AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I BOUGHT FOR AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT star index 2 gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt BOUGHT FOR DOLLARS lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star index 2 gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star 72 index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt l BOUGHT FOR DOLLARS pattern template srai DEBIT ACCT wallet AMT star index 2 gt DATE date NUMBER none PAYEE unknown CATEGORY lt star gt lt srai gt lt template gt lt category gt category pattern l BOUGHT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT star index 2 gt DATE da
11. gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT MY CARD FOR pattern template srai DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD FOR pattern template srai DEBIT ACCT star index 2 gt AMT lt star gt DATE date NUMBER none PAYEE star index 3 gt CATEGORY star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED MY CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt st
12. however that is defined before the learning machine group even defined its architecture Even faster than a child can learn language Here is another scenario Suppose we had a botmaster accelerator tool based on a kind of personality test The botmaster answers questions about his her personality background and preferences for a couple of hours Then the program classifies the botmaster into one of say 64 basic personality types Using the basic personality profile for that type the entire AIML set is automatically generated This may not be an exact digitization of the botmaster but it may be close enough for a first pass or say 6096 accurate the same as a personality test By applying the individual s biographical background and preferences the bot is not simply a clone of one of the 64 personality types but a unique individual personality The stark implication here is that theoretically we may be able to reduce the entire process of language learning and personality development to a couple of hours with a computer Consider the Botmaster Accelerator tool We can speed up the personality digitization process by administering a personality test and then selecting one of nine to twelve or sixteen say prototype bot personalities rather than editing the entire ALICE brain or starting from scratch Then using bot properties we can instantiate the bot with biographical data and preferences Now it doesn t rea
13. FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON CARD ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star 75 index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD ON FOR pattern template srai DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 4 gt lt srai gt lt template gt lt category gt category pattern l CHARGED DOLLARS ON MY CARD ON pattern template srai DEBIT ACCT star index 2 gt AMT lt star gt DATE date NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category CHARGED ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt DOLLARS ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT l
14. gt category gt lt pattern gt I DOLLARS ON MY FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt PUT ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY ATM FOR pattern template srai DEBIT ACCT checking AMT star DATE date NUMBER none PAYEE unknown CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT
15. lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED ON ATM CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED ON MY ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY 68 CARD pattern template srai DEBIT ACCT checking AMT lt star gt DATE lt date gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT DOLLARS ON MY ATM pattern template srai DEBIT ACCT checking AMT lt star gt DATE date srai template category gt category pattern l PUT ON MY CARD pattern template srai DEBIT ACCT checking AMT star DATE date srai template category gt category pattern l PUT ON MY ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE date srai template category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM CARDx pattern template srai DEBIT ACCT checking AMT star DATE lt date gt lt srai gt lt template gt lt category gt category gt
16. none unknown unknown none localhost localdomain DEBIT checking 1410 tue dec 25 08 34 07 pst 2001 number 101 Mr bill rent none localhost localdomain DEBIT visa 120 tue dec 25 08 35 00 pst 2001 none unknown the hotel none localhost localdomain DEBIT wallet 10 tue dec 25 08 35 28 pst 2001 none unknown a book none localhost localdomain THE AIML CATEGORIES The dialogue in section 2 resulted from the contents of the ALICE brain plus the following set of spreadsheet categories The basic idea is to reduce every financial statement into a transaction of the form CREDIT DEBIT or INITIALIZE There are two types of accounts CREDT accounts mainly for credit cards and BANK accounts which can include wallets piggybanks safes and other money containers as well as regular bank debit accounts like checking and savings The purpose of the INITIAIZE statement is to set an initial balance open a new account or override the spreadsheets calculated balance for an account CREDT accounts are opened with a negative balance initiated with a DEBIT transaction BANK accounts are opened with a positive balance initiated using a CREDIT transaction 60 All of the DEBIT and CREDIT transactions are recorded in the default gossip file This may not be the best place but it made for a simpler pedagogical example The program balance sh calculates the account balances as need see next section The following AIML
17. or reject it If the template contains a lt srai gt INTERJECTION lt srai gt then the pattern was either YES NO SO or one of a number of other interjections Usually these cases are the client saying YES or NO to a question the robot asked Here the botmaster may add a sensible response to the client s YES or NO reply so this makes a good target If the template contains a lt srai gt XFIND lt srai gt then the pattern was either WHAT IS WHO IS WHERE IS or some other similar default pattern for an information question These are usually good simple targets because the botmaster can often look up the answer in a dictionary or reference book if s he does not know the answer already But if the template has any other lt sr gt or lt srai gt this category is probably not a good source for targets Most cases of lt srai gt reduce complex forms of inputs into simpler ones The pattern DO YOU KNOW WHAT IS for example always reduces to lt srai gt WHAT IS star srai It is usually better to look for targets in the terminal patterns like WHAT IS that to write many specific new patterns based on reducible patterns like DO YOU KNOW WHAT IS The A L C E brain uses a special category with the pattern XFIND to respond to information questions for which she does not have a specific answer The XFIND category by itself does not produce useful targets so we disregard targets from this category even
18. outline a territory of language that has been effectively conquered by A L I C E and AIML No other theory of natural language processing can better explain or reproduce the results within our territory You don t need a complex theory of learning neural nets or cognitive models to explain how to chat within the limits of A L I C E s 40 000 categories Our stimulus response model is as good a theory as any other for these cases and certainly the simplest If there is any room left for higher natural language theories it lies outside the map of the A L 1 C E brain Academics are fond of concocting riddles and linguistic paradoxes that supposedly show how difficult the natural language problem is John saw the mountains flying over Zurich or Fruit flies like a banana reveal the ambiguity of language and the limits of an A L I C E style approach though not these particular examples of course A L 1 C E already knows about them In the years to come we will only advance the frontier further The basic outline of the spiral graph may look much the same for we have found all of the big trees from A to YOUR These trees may become bigger but unless language itself changes we won t find any more big trees except of course in foreign languages The work of those seeking to explain natural language in terms of something more complex than stimulus response will take place beyond our frontier increasing
19. predicates defining family relationships parent X X fail parent X Y mother X Y parent X Y father X Y grandmother X Y mother X Z parent Z Y Or we might consider more complex predicates such as 49 between X Y Y fail between X Y 2 between X Z Y and more complex queries such as What is a mammal Give me an example of a carnivore Name some pets Is Pittsburgh between New York and Chicago My mother is Mary Her mother is Millie Who is my grandmother In the end however it remains unclear whether interfacing A L 1 C E and AIML to a commonsense reasoning engine provides any real practical advantages over a pure AIML pattern targeting approach Instead of using Prolog to derive answers over a finite domain of terms we could just as well explicitly store all the answers in AIML category lt pattern gt IS A DOG A MAMMAL lt pattern gt template Yes template category gt category lt pattern gt IS DOG BIRD lt pattern gt lt template gt No lt template gt lt category gt lt category gt lt pattern gt IS A DOG ALIVE lt pattern gt lt template gt Yes lt template gt lt category gt and so on The number of combinations of all terms may produce a large number of combinations possibly requiring more memory than the AIML interpreter provides But even these patters have their own Zipf distribution
20. that In step 1 the patterns with match first because they are last in alphabetical order The order of the matches depends on this alphabetical ordering of patterns ALICE always matches suffixes with before prefixes with Whatever matches either wild card symbol becomes the value of lt star gt Steps 1 through 3 illustrate the common AIML templates that use the abbreviated sr tag Remember sr srai star srai The categories with the patterns and YOU MAY simply reduce the sentence to whatever matches the as illustrated by steps 1 and 3 Some AIML templates in ALICE combine the sr with an ordinary text response as step 2 with the pattern AGAIN The phrase Once more becomes part of any reply ending in AGAIN 23 The category in step 4 with SAY is a default that often produces logically correct but amusing dialogue Client Say hello in Swedish Robot Hello in Swedish or as in this case Client Say that Robot that Many patterns one reply The most common use of lt srai gt is to map two or more patterns to the same response category pattern P pattern template lt srai gt Q lt srai gt lt template gt lt category gt lt category gt lt pattern gt Q lt pattern gt lt template gt R lt template gt lt category gt An input matching either pattern P or Q gets the same response R To show
21. 59 EET ATVs CATEGORIES sa aset tur 60 THE SHEBESCRIELT itaqoe ca Ox mot A RON APO I SR Na Ng e NE 63 AN AIML SPREADSHBEET ASS due M ade u 63 Appendix A AIML categories for personal 64 CHAPTER VIII Building a Subscriber Based Bot Business with Pandorabots 77 Ten Killer Apps for L E and AIMLI 71 CHAPTER IX The Euture or AIMD a naat 80 CONCLUDING REMARKS 82 REFERENCES anaq t u E s u a y S ANO 83 ACKNOWLEDGEMENTS LIL RE A RARE 84 COMMENTS ABOUT THE AUTHQOR REI Se Pepe Vae 84 INTRODUCTION If you are looking for a revolution in IT you ve come to the right place There are several ways in which A L C E and AIML defy the dominant computer science paradigms For one thing the ways in which humans and computers normally communicate are very different Humans tend to spend a lot of time on chit chat and informal dialogue with little or no purpose Computers are known for giving precise true and logical answers The rate of information exchange in most human dialogue is very low no more than 1Kbit per second but computer communication
22. AIML It s called the brain Knowing only a few details of neuroscience it is actually easy to imagine how the brain could implement a stimulus response category It is also easy to imagine the brain executing a n lt srai gt through a feedback circuit It is much harder to understand how the brain would implement condition or javascript Keeping AIML pure enough so that one day some scientist smarter than us might be able to map our knowledge onto the human brain is an exciting prospect for the Reductionist approach Can we have it both ways The answer is yes think if we are careful to specify the interfaces to AIML The system and javascript tags already provide two ways to process information outside the AIML engine In the future we should consider more carefully how to make ALICE a container that is easily accessed by other applications Another solution is provided by namespaces which allow embedding of other XML languages in AIML The Reductionists will be satisfied so long as there are few changes to the AIML tagset itself 11 CHAPTER l Overview of AIML AIML or Artificial Intelligence Mark up Language enables people to input knowledge into chat bots based on the A L I C E free software technology AIML was developed by the Alicebot free software community and during 1995 2000 It was originally adapted from a non XML grammar also called AIML and formed the basis for the first Alicebot A
23. MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT DOLLARS ON MY ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT DOLLARS ON MY ATM FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY ATM ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY ATM CARD FOR lt pattern
24. a more concrete example the input Hello should have an appropriate response like Hi there But we can expand the inputs generating this response to include all the common variations of Hello lt category gt lt pattern gt Hl lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt lt category gt lt pattern gt HOWDY lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt lt category gt lt pattern gt HALLO lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt lt category gt lt pattern gt HI THERE lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt 34 category pattern HELLO pattern lt template gt Hi there lt template gt lt category gt As the following example shows we can use the lt sr gt tag as an abbreviation for lt srai gt lt star gt lt srai gt This category creates a compound response to both hello and whatever matches lt category gt lt pattern gt HELLO lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt sr gt lt template gt lt category gt MULTIPLE WILDCARDS Some AIML authors may ask why AIML does not yet allow two or more wildcard symbols per pattern In the present day Java implementation the symbols and appear at most once in the AIML
25. an S shape that looks like exponential growth at first but then eventually levels off as the technology saturates the marketplace Television is an instructive example Introduced commercially in 1947 growth was slow until the knee in 1949 51 Following that TV spread at an approximately linear rate until about 1958 when it began to slowly level off By 1980 nearly 10095 of households had TV sets Philo T Farnsworth like myself a native of Maine and resident of San Francisco invented the cathode ray picture tube here in 1927 He had to wait over twenty years before his invention saw the commercial light of day Another interesting trend is the increasing steepness of the curves The telephone introduced in 1876 did not saturate the market until about 100 years later The curve for AM radio starts in 1923 and levels off in 1963 a span of about 40 years The VCR was launched in 1980 but levels off in 2000 only about twenty years Of course there are a lot of other unfinished curves crowded near the end of the graph cell phone PC internet camcorder etc whose future courses remain uncertain 82 Another analysis might categorize the curves into those that look more like an S curve and those that look more like a straight slope The curve for the telephone is actually a bit ragged taking a big dip downward during the Great Depression from 4096 to 30 of the market the same cannot be said of any other technology on the c
26. categories represent a first pass attempt to create the simple spreadsheet Not a true double entry bookkeeping System the categories nonetheless demonstrate the feasibility of building a personal finance application in AIML category TRANSACTION lt pattern gt template random li Ok li li Yes li lt li gt Alright lt li gt lt li gt l got lt random gt lt random gt 315 there anything else can help you with lt li gt lt li gt Do you have more transactions lt li gt 3 have recorded your transaction lt li gt lt li gt Anything else can do for you lt li gt li Thanks for the information li li Is that it lt li gt lt random gt lt template gt lt category gt lt category gt lt pattern gt CREDIT ACCT AMT DATE NUMBER PAYER CATEGORY NOTE lt pattern gt lt template gt srai ACKNOWLEDGE TRANSACTION lt srai gt lt think gt lt gossip gt CREDIT lt star gt lt star index 2 gt lt star index 3 gt lt star index 4 gt lt star index 5 gt lt star index 6 gt lt star index 7 gt lt id gt lt gossip gt lt think gt lt template gt lt category gt lt category gt lt pattern gt DEBIT ACCT AMT DATE NUMBER PAYEE CATEGORY NOTE lt pattern gt lt template gt lt srai gt ACKNOWLEDGE TRANSACTION lt sra
27. category gt category gt lt gt 1 CHARGED ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt gt 1 DOLLARS ON CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 4 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4
28. category gt category gt lt pattern gt WHAT IS MY CARD BALANCE lt pattern gt lt template gt lt srai gt BALANCE ACCT lt star gt lt srai gt lt template gt lt category gt lt category gt lt pattern gt DEPOSIT IN CHECKING 62 ACCOUNT pattern template srai CREDIT ACCT checking AMT star srai template category gt category gt lt gt PAID MY RENT WITH CHECK pattern template srai DEBIT ACCT checking AMT get name rent gt DATE lt date gt NUMBER lt star gt PAYEE get name landlord gt CATEGORY RENT lt srai gt lt template gt lt category gt category gt lt pattern gt I PAID THE RENT WITH CHECK pattern template srai DEBIT ACCT checking get gt DATE lt date gt NUMBER lt star gt PAYEE get name landlord gt CATEGORY RENT lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt lt aiml gt The remainder of the AIML categories have been placed in an appendix due to considerations of length THE SHELL SCRIPT There are an infinite number programs one could write to calculate account balances based on the DEBIT and CREDIT
29. is much faster Alicebot AIML is an attempt to bridge this divide We have dispensed with much of the conventional wisdom from structured programming too Where less code is usually good more AIML is usually better Filling up RAM with a lot of code that is seldom activated is a good idea in AIML Allowing a novice programmer to write a large and unwieldy program in AIML is also unconventionally good practice Not to mention the basic minimalist approach of AIML is borrowed from ELIZA a design that was largely abandoned by the research establishment indeed even derided as a toy Yet the success of A L I C E and AIML if anything put the research establishment to shame for their years of research and millions of spending they have achieved nothing close to the performance of A L I C E in natural language processing Even the success stories of Al research such as the chess playing supercomputers required far higher expenditures than A L C E and AIML The conventional model of an information service assumes that a computer will always provide accurate replies promptly People have asked what is the difference between Ask Jeeves and A L C E Jeeves is designed so that the client will ideally ask one question and then immediately receive an accurate reply including a hyperlink to another site After just one transaction the client clicks on the link and leaves the Ask com site A L I C E on the othe
30. it merely has to find one ISA HIERARCHY The first step is to construct an Isa hierarchy in AIML The simplest entries look like this category gt lt pattern gt ISA RAVEN lt pattern gt lt template gt A Bird lt template gt lt category gt category lt pattern gt ISA BIRD lt pattern gt 52 lt template gt An animal lt template gt lt category gt Basically the AIML expression is equivalent to the assertion Isa raven bird But the Isa function is one to many For example chicken is both a bird and a food We use the AIML lt random gt tag to group these assertions lt category gt lt pattern gt ISA CHICKEN lt pattern gt lt template gt lt random gt lt li gt A Bird lt li gt lt li gt A Food lt li gt lt random gt lt template gt lt category gt Finally the AIML set should provide a default case for the Isa relation in case the first argument can not be identified category lt pattern gt ISA lt pattern gt lt template gt Unknown lt template gt lt category gt KNOWLEDGE BASE The second step was to program the knowledge base with Has and Does relations It is not necessary to use an abbreviated predicate name in an AIML pattern so we can actually encode the knowledge base quite conveniently in natural language So rather than write does bird fly we can express the relation in AIML as category gt pattern WHAT DOES A BIRD DO
31. like to meet the man think the man is rather smart either he s got us all thinking is ALICE or he s actually got us thinking ALICE is him Either way he s won have much more respect for Wallace after reading this reply He s a deeply insightful individual and doesn t appear to be taken in by much of the bullshit of the AI field Goddamn what a thoughtful set of paragraphs This is the first slashdot article l ve decided to print don t care about the length who knew an specialist could be such a skilled writer amazing interview The insights on Al particularly the digression into the functions of AIML for A L I C E were wonderful in this interview 85 Index A L I C E 5 12 19 20 21 41 78 80 83 ALICE A I Foundation 80 84 85 Ask Jeeves 5 10 atomic patterns 13 bot 2 15 16 17 37 40 45 47 49 58 59 77 78 79 80 81 82 botmaster 2 3 13 16 17 35 37 38 40 41 42 43 44 78 79 81 82 Botmaster Control page 79 category 11 12 13 14 15 16 17 18 19 21 34 35 36 37 38 39 41 42 43 44 48 50 52 53 54 55 56 57 59 61 62 63 64 65 66 67 68 69 70 71 72 73 74 15 76 77 client 5 15 16 18 33 35 36 41 43 44 48 55 56 59 Conditionals 13 CYC 45 Divide and Conquer 13 ELIZA 5 9 15 38 Experimentalist 10 Kawasaki 82 83 keywords 13 learning 10 16 17 20 31 38 81 Linux 5 45 mat
32. lt pattern gt lt template gt Fly lt template gt lt category gt The fact that not all birds fly should not distract you from understanding this example As was the case with the Isa relation the Has and Does relations are many to one We again use the lt random gt tag to group the many assertions together lt category gt lt pattern gt WHAT DOES A BIRD HAVE lt pattern gt lt template gt lt random gt lt li gt Lungs lt li gt 53 lt li gt An lt gt lt li gt A beak lt li gt lt li gt A tail lt li gt lt li gt A wing lt li gt lt li gt A feather lt li gt lt li gt Cold blood lt li gt lt random gt lt template gt lt category gt DEFAULTS Some default replies are also necessary to prevent infinite loops Here the Art of AIML writing comes into play The ultimate default Isa relation returns Unknown so the symbol UNKNOWN appears in the patters of these defaults category gt lt pattern gt WHAT DOES UNKNOWN lt gt template random lt li gt lmagine no possessions li lt li gt don t know lt li gt lt li gt The same as everyone else lt li gt lt random gt lt template gt lt category gt lt category gt pattern WHAT DOES UNKNOWN DO lt pattern gt template random lt li gt Exist lt li gt lt li gt don t know li lt li gt The same as everyone else lt li gt lt
33. new atomic pattern WHAT IS A PIZZA The botmaster now needs only to click the lt srai gt button and add some text to create the template lt srai gt WHAT IS PIZZA lt srai gt since we previously added that category 3 In the third example we start with the following dialogue fragment Client My name is Rich Robot Rich are you my master Client Yes Robot Oh The topic is READY The botmaster views the display Input YES RICH ARE YOU MY MASTER READY Matched YES New AIML YES Here the crude targeting algorithm fails to create a new target 43 pattern but it has nevertheless detected a good new target The botmaster edits the New AIML text boxes so that they contain New AIML YES ARE YOU MY MASTER Now s he edits the template to read Okay will obey only you And then saves the new category This example shows how targeting can be used to detect new that categories 4 In the next example the client asks WHAT IS THE IRON GIANT The value of that is PLEASE STAND BY and the topic is PIZZA The Targeting display shows Input WHAT IS THE IRON GIANT PLEASE STAND BY PIZZA Matched WHAT IS THE New AIML WHAT IS THE IRON In this case the algorithm extended the pattern by one word but since the input still contained more words the program made the new AIML pattern end with the wildcard This new pattern is not quite perfect because the question
34. one matched category but also any others recursively reached through srai RECURSION AIML implements recursion with the lt srai gt operator No agreement exists about the meaning of the acronym The stands for artificial intelligence but S R may mean stimulus response syntactic rewrite symbolic reduction simple recursion or synonym resolution The disagreement over the acronym reflects the variety of applications for lt srai gt in AIML Each of these is described in more detail in a subsection below 1 Symbolic Reduction Reduce complex grammatical forms to simpler ones 2 Divide and Conquer Split an input into two or more subparts and combine the responses to each 3 Synonyms Map different ways of saying the same thing to the same reply 4 Spelling or grammar corrections b Detecting keywords anywhere in the input 6 Conditionals Certain forms of branching may be implemented with lt srai gt 7 Any combination of 1 6 The danger of srai is that it permits the botmaster to create infinite loops Though posing some risk to novice programmers we surmised that including srai was much simpler than any of the iterative block structured control tags that might have replaced it SYMBOLIC REDUCTIONS Symbolic reduction refers to the process of simplifying complex grammatical forms into simpler ones Usually the atomic patterns in categories storing robot know
35. pattern One would sometimes like to write patterns like lt pattern gt IS COOL lt pattern gt Wildcard patterns beginning and ending with are an especially attractive subset to consider At the present time the ALICE program classifies all the inputs matching patterns like lt pattern gt IS COOL lt pattern gt into the default category stored in Pickup aiml By analyzing the inputs classified as default the botmaster finds common suffixes The file Suffixes aiml contains a fairly large number of categories like this one lt category gt lt pattern gt _ DO NOT YOU THINK lt pattern gt lt template gt lt srai gt do you think lt star gt lt srai gt lt template gt lt category gt These categories serve to rewrite the input in a simpler form more likely to match another simple pattern In this case the recursive response might match lt pattern gt DO YOU THINK lt pattern gt Symbolic reductions and state The tag lt that gt in AIML introduces one dimension of dialogue state into the robot response The value of lt that gt is whatever the robot said before that provoked the current client input The inputs yes and no are two of the most common human queries But a careful analysis of the dialogues shows that most of the time people say yes or no to only a small set of questions asked by the robot Our file YesNo aiml contains the most common categories activated by yes or no The categor
36. pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt WITHDRAW FROM MY SAVINGS ACCOUNT pattern template srai DEBIT ACCT savings AMT star srai template category gt category pattern WITHDRAW FROM SAVINGS ACCOUNT pattern template srai DEBIT ACCT savings AMT star srai template category category pattern WITHDRAW FROM SAVINGS pattern template srai DEBIT ACCT savings AMT lt star gt lt srai gt lt template gt lt category gt category WITHDREW FROM SAVINGS pattern template srai DEBIT 69 ACCT savings AMT lt star gt lt srai gt template category gt category pattern DEBIT ACCT AMT lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star gt AMT star index 2 gt DATE lt date gt lt srai gt lt template gt lt category gt category pattern DEBIT ACCT AMT DATE lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star gt AMT star index 2 gt DATE star index 3 gt NUMBER NONE lt srai gt lt template gt lt category gt category pattern DEBIT ACCT AMT DATE NUMBER PAYEE CATEGORY lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star gt AMT star index 2 gt DATE lt star index 3 gt NUMBER lt star index 4 gt PAYEE l
37. yet discovered the true potential of virtual personalities and bots and our mission is to design tools which will help Bots achieve greater relevance in both the commercial and personal sectors Capabilities such as voice recognition text to speech synthesis invoking programs on remote or local machines content development tools e commerce and cartoon animations are under development to help you bring your bot to virtual life You should expect numerous changes to this site so please save your bot creations on your own machine using the file save facility of your browser Your bot s knowledge is typically stored in a file called update aiml and is available via the edit link which is itself available from the Botmaster control link Rudimentary documentation now exists see the Help link and will undergo changes as the site changes The Bots hosted here conform in most respects to the bot specifications given at http www alicebot org The Subscriber Service allows you to receive money from customers interacting with your robot at www pandorabots com Here s now it works Using the hosting service www pandorabots com you create and publish a robot and you don t reveal the published location to anyone You recruit and persuade customers to sign up to interact with your robot You decide who talks to your robot the amount to charge and how to collect the money Click the Subscriber services button and follow the directions to allo
38. 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ATM AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt temp
39. ENT ON CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED DOLLARS ON MY FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star 74 index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON MY CARD FOR ON pattern template srai DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt l CHARGED ON MY ATM FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT DOLLARS ON MY ATM CARD FOR ON lt pa
40. I call the Reductionist and the Experimentalist factions The Reductionists belong to the long tradition of mathematical logic that began with Aristotle They would keep the language as simple as possible if not simpler so that it could withstand the rigor of mathematical analysis theory and proof Reductionists are strongly attracted to the minimalist design philosophy of AIML Experimentalists are those who see the minimal AIML that exists today as the mere seed of a great tree of a new language There is no shortage of ideas for adding new features and tags to AIML in support of peer to peer learning voice recognition character animation robot control general purpose programming language features and many more These all take AIML beyond the simplistic language favored by the Reductionists In some sense the Experimentalists are driven by the need to create practical applications with AIML technology and the perception that additional features will enable those applications An early version of this debate appeared when raised the question Is AIML a high level or a low level language Tom Ringate pointed out that low level means close to the machine in terms of hardware and high level languages are abstracted far from the details of the underlying architecture By that reasoning AIML is a high level language like Lisp C Java or SETL 10 But said wait a minute there is a hardware model for
41. IBANK BALANCE IS lt pattern gt lt template gt lt srai gt INITIALIZE CREDT ACCT Citibank AMT lt star gt lt srai gt lt template gt lt category gt category gt lt gt SAVINGS ACCOUNT BALANCE IS lt pattern gt lt template gt lt srai gt INITIALIZE BANK ACCT savings AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt MY VISA CARD BALANCE IS lt pattern gt lt template gt lt srai gt INITIALIZE CREDT ACCT visa card AMT star srai template category gt category pattern l HAVE MY WALLET lt pattern gt lt template gt lt srai gt INITIALIZE BANK ACCT wallet AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I HAVE DOLLARS IN MY WALLET lt pattern gt lt template gt lt srai gt INITIALIZE BANK ACCT wallet AMT lt star gt lt srai gt lt template gt lt category gt lt category gt lt pattern gt WHAT IS MY ACCOUNT BALANCE lt pattern gt lt template gt lt srai gt BALANCE ACCT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I HAVE DOLLARS TO MY NAME lt pattern gt lt template gt lt srai gt INITIALIZE BANK ACCT wallet AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I HAVE DOLLARS lt pattern gt lt template gt lt srai gt INITIALIZE BANK ACCT wallet AMT lt star gt lt srai gt lt template gt lt
42. Java or C Thus we can query the psych program from the shell with the command 96 echo isa dog animal psych which will print some text including the Prolog reply Yes or No The obvious system command would be something like lt system gt echo isa star index 1 gt lt star 2 gt psych lt system gt Unfortunately the implementation of lt system gt in programs B D and dB uses the Java Runtime exec method The exec method has some known problems dealing with Unix pipes and the echo command THE SHELL SCRIPT To get around the problems with pipes and echo in Java Runtime exec an intermediate shell script was created This script called isa sh contains the line echo isaset 1 2 psych tail 2 head 1 The echo command inserts the arguments to isa sh 1 and 2 into the Prolog query string isaset 1 2 For example if we typed the command isa sh dog animal The echo command produces the string isaset dog animal The psych command executes the prolog program The final two commands in the pipeline tail and head simply strip off excess output lines generated by Prolog THE PROLOG PROGRAM Another minor stumbling block in this experiment turned out to 46 be that Prolog does not implement isa hierarchies very cleanly Nevertheless the program PSYCH implements a small isa hierarchy well enough to answer some basic yes no questions An i
43. L I C E the Artificial Linguistic Internet Computer Entity AIML describes a class of data objects called AIML objects and partially describes the behavior of computer programs that process them AIML objects are made up of units called topics and categories which contain either parsed or unparsed data Parsed data is made up of characters some of which form character data and some of which form AIML elements AIML elements encapsulate the stimulus response knowledge contained in the document Character data within these elements is sometimes parsed by an AIML interpreter and sometimes left unparsed for later processing by a Responder CATEGORIES The basic unit of knowledge in AIML is called a category Each category consists of an input question an output answer and an optional context The question or stimulus is called the pattern The answer or response is called the template The two types of optional context are called that and topic The AIML pattern language is simple consisting only of words spaces and the wildcard symbols and The words may consist of letters and numerals but no other characters The pattern language is case invariant Words are separated by a single space and the wildcard characters function like words The first versions of AIML allowed only one wildcard character per pattern The AIML 1 01 standard permits multiple wildcards in each pattern but the language is designed to be a
44. MBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY ATM CARD AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt 66 NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE date NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ATM CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt
45. NO The most common multiple word input is HOW OLD ARE YOU This type of analysis which cost Dr Zipf many hours of labor is now accomplished in a few milliseconds of computer time Considering the vast size of the set of things people could possibly say that are grammatically correct or semantically meaningful the number of things people actually do say is surprisingly small Steven Pinker in his book How the Mind Works wrote that Say you have ten choices for the first word to begin a sentence ten choices for the second word yielding 100 two word beginnings ten choices for the third word yielding a thousand three word beginnings and so in Ten is in fact the approximate geometric mean of the number of word choices available at each point in assembling a grammatical and sensible sentence A little arithmetic shows that the number of sentences of 20 words or less not an unusual length is about 1020 Fortunately for chat robot programmers Pinker s combinatorics are way off Our experiments with ALICE indicate that the number of choices for the first word is more than ten but it is only about two thousand Specifically 1800 words covers 95 of all the first words input to ALICE The number of choices for the second word is only about two To be sure there are some first words I and You for example that have many possible second words but the overall average is just under two words The average branching factor d
46. NOWLEDGEMENTS This research was funded by donations to and carried out by volunteers of the ALICE Foundation Inc nonprofit charitable research corporation The author is grateful to Peter Plantec for his careful reading and comments on an early draft of this manuscript COMMENTS ABOUT THE AUTHOR Some of the best comments about the Slashdot Interview with Dr Wallace on http slashdot org interviews At the time it appeared it was the longest single article published in the history of slashdot and had to be broken into three parts If you can t find a useful quote in this interview to use as a sig you re weak really weak this is the BEST interview I ve ever read on bar none there is some serious doubt on this very forum whether this is ALICE or the good Doctor He does a good job off coming off as a troll without his very own impersonator This was one of the best interviews l ve ever read This guy is a genius this man is brilliant and only wish he had written more 84 Great interview Probably the best l ve ever read on Slashdot and I ll definitely come back eventually to read everything glazed over Does anyone else think it s strange that the leading researcher in the world is a self described mental patient This is honestly one of the best interviews or literary pieces have ever read He is one of the most though provoking people I ve read and I d honestly
47. Some of them are much more likely to be asked than others In this respect one can imagine an AIML knowledge base not quite equivalent to the output of a Prolog program but close enough for all practical purposes PROLOG ASSERTIONS In addition to the associative rules for the isa hierarchy the PSYCH program includes the following basic commonsense assertions isa human mammal isa man human isa woman human isa baby human 50 isa boy human isa girl human isa rich man isa carnivore mammal isa cat carnivore isa dog carnivore isa mammal animal isa animal alive isa plant alive isa fruit plant isa tree plant isa fir tree isa maple tree isa spruce tree isa birch tree isa grass plant isa fruit food a weed plant isa vegetable plant isa vegetable food isa alive entity isa rock entity isa sand rock isa alien alive isa cat pet isa dog pet p G gt gt nuu wu isa chicken food isa chicken bird isa bird animal j isa duck bird isa duck food isa penguin bird isa cow cattle isa bull cattle isa cattle herbivore isa cattle food isa herbivore animal isa car vehicle isa plane vehicle isa jet plane isa train vehicle isa bart t
48. T DOLLARS ON MY ATM CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY CARD ON pattern template srai DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ON pattern template DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SP
49. VORITE COLOR 12 WHAT IS 2 PLUS 2 12 WHAT IS A CAR 12 WHAT IS A DOG 12 WHAT IS ARTIFICIAL INTELLIGENCE 12 WHAT IS IT ABOUT 12 WHAT IS LIFE 12 WHAT IS SEEKER 12 WHAT IS YOU NAME 12 WHAT IS YOUR FAVORITE 12 WHAT IS YOUR SURNAME 11 WHAT IS 1 1 11 WHAT IS A CHATTERBOT 11 WHAT IS A PRIORI 11 WHAT IS SETL 11 WHAT IS THE TIME IN USA 11 WHAT IS THE WEATHER LIKE THERE 11 WHAT IS YOUR FAVORITE FILM 10 WHAT IS A CATEGORY C CLIENT 10 WHAT IS A PENIS 10 WHAT IS BOTMASTER 10 WHAT IS MY IP ADDRESS 10 WHAT IS THE DATE 10 WHAT IS THIS 10 WHAT IS YOUR ADDRESS 10 WHAT IS YOUR FAVORITE MUSIC 10 WHAT IS YOUR FAVORITE OPERA 10 WHAT IS YOUR GOAL 10 WHAT IS YOUR IP ADDRESS Even subsets of natural language like the example shown here of sentences starting with WHAT IS tend to have Zipf like distributions Natural language search bots like Ask Jeeves are based on pre programmed responses to the most common types of search questions people ask A REVOLUTION IN A l One sure sign of a successful revolution is when the original revolutionaries break into factions and start squabbling among themselves Such debate is a luxury before the status quo is overthrown but an almost necessary feature for stability afterward The Alicebot and AIML free software movement has that kind of revolutionary success in spades judging from the amount of squabbling that has gone on Broadly the AIML community has divided into two camps which
50. WHAT IS THE IRON is too general So the botmaster may choose to edit the display so that it shows New AIML WHAT IS THE IRON GIANT Then write a template and save the result b In this example the botmaster sees the display Input OH SEE YOU LATER YOU Matched OH New AIML OH This information is almost useless so the botmaster discards the target with the delete target button This ensures that the target will not reappear when browsing with next target 6 Now suppose the value of that is TELL ME and the topic is A RIDDLE The client enters COULD POSSIBLY GIVE YOU A HINT targeting display shows Input COULD POSSIBLY GIVE YOU A HINT TELL ME A RIDDLE Matched COULD New COULD POSSIBLY The botmaster recognizes that the word POSSIBLY plays little role in resolving the client s meaning so s he uses the Reduce button to create the template srai l COULD star srai The effect of the new category is to eliminate the word POSSIBLY 44 from all inputs beginning with I COULD POSSIBLY CHAPTER V PSYCH Activating Prolog from AIML This chapter describes a first experiment linking an Artificial Intelligence Markup Language AIML script to a Prolog program Specifically we used GNU gprolog to implement a program called PSYCH pronounced CYC The AIML script uses the system command to activate the PSYCH commonsense reasoni
51. an Try asking me again User Where do find the web page with Australopithecus Alice http www ancientskulls net australopithecus html This is a very nice example This is exactly the same as targeting when the botmaster is the teacher Or should say it is an AIML interface to targeting Of course this scenario raises many problems such as how to weed out hackers and other garbage givers which might mean only turning on the Learn function while in a restricted sandbox though it is the free labor of many chatters the web that could help the bot grow wise quickly That is the crux of the learning problem in a nutshell It all comes down to editing Either the botmaster edits good quality replies to begin with or if the bot is open to the public then the teacher has to edit out all the crappy answers supplied by all that free labor My personal view is that compared with a hypothetical learning machine capable of learning language like a child the AIML targeting approach is more efficient CONTEXT The keyword that in AIML refers to the robot s previous utterance Specifically if the robot responds with a multiple sentence paragraph the value of that is set to the last sentence in the sequence The choice of the keyword that is motivated by its use in ordinary language R Today is yesterday C That makes no sense R The answer is 3 1412926 approximately C That is cool In AIML the syntax lt that gt
52. ar DATE lt date gt lt srai gt lt template gt lt category gt category pattern l WITHDREW FROM MY lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt srai template category gt category gt lt gt 1 WITHDREW FROM lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt lt star gt lt srai gt lt template gt lt category gt category gt lt gt PAID DOLLARS FOR AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE date NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT star DATE date NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS AT ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE date NUMBER none PAYEE star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PAID DOLLARS FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template
53. ar gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt DOLLARS ON MY CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY CARD AT lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt gt DATE lt date gt NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD AT lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE 70 date NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD AT pattern template srai DEBIT ACCT star index 2 gt lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt gt 1 CHARGED DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt
54. ate gt NUMBER lt star gt PAYEE star index 2 gt lt srai gt lt template gt lt category gt lt category gt lt pattern gt WITHDRAW FROM CHECKING ACCOUNT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt WITHDRAW FROM CHECKING lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt lt category gt lt pattern gt WITHDRAW FROM MY CHECKING ACCOUNT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I GOT FROM THE ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I WITHDREW FROM THE ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I TOOK OUT FROM THE ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I TOOK OUT WITH THE ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I GOT OUT WITH THE ATM lt
55. ation along these lines would use some other symbols to serve the same function 22 You can see that the matching algorithm specifies an effective procedure for searching the filesystem for a particular file called template txt The path name distinguishes all the different template txt files from each other What s more you can visualize the compression of the Graphmaster in the file system hierarchy All the patterns with common prefixes become compressed into single pathways from the root Clearly this storage method scales better than a simple linear array or database storage of patterns whether they are stored in RAM or on disk 3 Graphmaster For Dummies Here is a really simple explanation of Graphmaster pattern matching It works just like a dictionary or encyclopedia If you want to look up a word or phrase you don t start at the beginning or the end and search through every entry until you find a match No you turn first to the section that matches the first letter or word Then you skip to another section that contains a set beginning with the next letter or word You continue in this process until you find your word or phrase In the Graphmaster the wildcards and act like two special letters that come before 0 and after Z respectively You should always check the listings first in case your input has an ending listed there Check the alphabetical listings next If you don t find any matches
56. attern gt WHAT DOES HAVE lt pattern gt lt template gt lt srai gt WHAT DOES srai lSA lt star gt lt srai gt HAVE lt srai gt lt template gt lt category gt Using these deductions we can now ask the robot What does a raven have and even though there is no explicit answer the program can infer that since a raven is a bird it must also have anything a bird has feathers beak tail eyes lungs wings and cold blood PLURALS A slight modification to the above categories allows us to handle plurals as well as singular forms in this class of questions If we replace the ISA string with ISB inside the lt srai gt tags above we can define an Isb relation as a gateway or filter on top of Isa The Isb function simply transforms its argument to a singular form and recursively applies Isa 55 category gt lt pattern gt ISB lt pattern gt lt template gt lt srai gt ISA lt srai gt SINGULAR lt star gt lt srai gt lt srai gt lt template gt lt category gt The knowledge base for the Singular relation is just like the category sets for Isa Does and Has lt category gt pattern SINGULAR WOLVES lt pattern gt lt template gt wolf lt template gt lt category gt lt category gt pattern SINGULAR WOMEN lt pattern gt lt template gt woman lt template gt lt category gt lt category gt lt pattern gt SINGULAR WORKS OF ART lt pattern gt lt template gt work
57. ching 13 21 22 23 33 34 35 38 39 40 Copyright 2003 ALICE A I Foundation Inc neuroscience 11 Pandorabots 4 77 78 79 80 pattern 6 12 13 14 15 16 17 18 19 21 22 23 24 33 34 35 36 37 38 39 40 41 42 43 44 48 49 50 52 53 54 55 56 57 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 PayPal 78 79 80 Pinker 8 19 Program dB 41 84 Prolog 3 45 46 47 48 49 50 52 58 83 84 PSYCH 3 45 47 48 49 50 reasoning 10 45 48 49 50 52 58 recursion 13 Reductionist 10 11 Spelling 13 Star Trek 6 64 stimulus response 11 12 13 20 32 38 symbolic reduction 13 14 33 Symbolic reduction 13 synonym resolution 13 Targeting 16 43 44 78 template 12 13 14 15 16 18 19 21 22 23 26 30 33 34 35 36 37 38 40 A1 42 43 A4 48 50 52 53 54 55 56 57 61 62 63 64 65 66 67 68 69 70 71 72 13 74 75 76 77 The Graduate 77 voice recognition 10 79 wildcard 12 14 19 22 36 42 44 WIMP 64 Windows 22 45 64 Zipf 6 Zipfs Law 6 84
58. d to handle requests for refunds or have a strict no refunds policy In the case of the ALICE A I Foundation subscriptions to the A L I C E Silver Edition are considered donation to a nonprofit charity and refunds are not given no exceptions You can use PayPal to refund the customer his or her money but they should remember that PayPal collects a fee for every transaction As part of the cost of doing business a full refund cannot be expected The buyer should beware that even a money back guarantee may not include all the fees required to process both the purchase and the refund through an e payment system like PayPal On Pandorabots the Subscriber page has a feature to deactivate subscribers for any reason In the first column marked State locate the link called Click to deactivate and follow the directions to remove a subscriber from your bot subscription list The easiest way to receive money from customers is through the service www PayPal com Other ways include checks cash and credit cards You can also use wire transfer services such as Western Union or postal money orders Receiving money Credit Cards is difficult unless you have a merchant credit card account and so we recommend you seek payment using PayPal Of course you can also provide free access to your robot CHAPTER IX The Future of AIML began working on a new bot named Interviewer that is designed to automate some of the botmastering proce
59. e robotics Mobile robots work best fastest and 38 demonstrate the most animated realistic behavior when their sensory inputs directly control the motor reactions Higher level symbolic processing search and planning tends to slow down the process too much for realistic applications even with the fastest control computers Why Is An AIML Category Called A Category The term was borrowed from pattern recognition theory The general pattern recognition problem is to partition a space P of inputs into disjoint regions so that the pattern classifier can categories a point x from P into one of those regions The regions are called categories C1 C2 Cn Formally the union of the Ci P and the intersection of any pair of Ci and Cj the empty set whenever i not equal to j The pattern recognition problem is to categorize x into one of the Ci In many cases the partition is define by a matching function f x i which computes a distance from x to the category Ci For any given point x in P x is categorized as Ci provided f x i not equal to f x j for by other category j If the input space consists of 32 digit bar code scans and the categories represent different items for sale then the problem is to classify a given 32 bit input into the nearest matching code for one of those items If the input space consists of 250 000 pixel TV pictures of human faces and the categories represent a set f wanted terrori
60. e PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE
61. e pattern There is only one giant Microsoft followed by a number of large and medium sized firms and then a large tail of small and very small firms Zipf was independently wealthy This is how he could afford to hire a room full of human computers to count words in newspapers and periodicals Each computer would arrive at work and begin tallying the words and phrases directed by Zipf These human computers found that Zipf s Law applies not only to words but also to phrases and whole sentences of language 8024 YES 5184 NO 2268 OK 2006 WHY 1145 BYE 1101 HOW OLD ARE YOU 946 HI 934 HOW ARE YOU 846 WHAT 840 HELLO 663 GOOD 645 WHY NOT 584 OH 553 REALLY 544 YOU 531 WHAT IS YOUR NAME 525 COOL 516 DO NOT KNOW 488 FUCK YOU 486 THANK YOU 416 SO 414 ME TOO 403 LOL 403 THANKS 381 NICE TO MEET YOU TOO 375 SORRY 374 ALICE 368 HI ALICE 366 OKAY 353 WELL 352 WHAT IS MY NAME 349 WHERE DO YOU LIVE 340 NOTHING 309 KNOW 303 WHO ARE YOU 300 NOPE 297 SHUT UP 296 LOVE YOU 288 SURE 286 HELLO ALICE 277 HOW 262 WHAT DO YOU MEAN 261 MAN 251 WOW 239 SMILE 233 ME 227 WHAT DO YOU LOOK LIKE 224 SEE 223 HA 218 HOW ARE YOU TODAY 217 GOODBYE 214 NO YOU DO NOT 203 DO YOU 201 WHERE ARE YOU The human input histogram ranking the number of times ALICE receives each input phrase over a period of time shows that human language is not random The most common inputs are YES and
62. e per second But a more realistic application may also have to include record identifiers The most sophisticated language processing demonstrated in this 59 experiment was the use of AIML predicates for landlord and rent This makes possible the transformation of sentences such as paid the rent with check number 101 into a transaction INITIALIZE BANK checking 2300 Tue Dec 25 08 32 16 PST 2001 localhost localdomain CREDIT checking 2300 tue dec 25 08 32 16 pst 2001 none unknown unknown none localhost localdomain INITIALIZE BANK savings 1000 Tue Dec 25 08 32 24 PST 2001 localhost localdomain CREDIT savings 1000 tue dec 25 08 32 24 pst 2001 none unknown unknown none localhost localdomain INITIALIZE CREDT visa card 2100 Tue Dec 25 08 32 32 PST 2001 localhost localdomain DEBIT visa card 2100 tue dec 25 08 32 32 pst 2001 none unknown unknown none localhost localdomain INITIALIZE CREDT Citibank 9000 Tue Dec 25 08 32 51 PST 2001 localhost localdomain DEBIT Citibank 9000 tue dec 25 08 32 51 pst 2001 none unknown unknown none localhost localdomain INITIALIZE BANK wallet 200 Tue Dec 25 08 33 01 PST 2001 localhost localdomain CREDIT wallet 200 tue dec 25 08 33 01 pst 2001 none unknown unknown none localhost localdomain DEBIT wallet 10 tue dec 25 08 33 23 pst 2001 none a taxi unknown none localhost localdomain DEBIT visa 23 for lunch tue dec 25 08 33 34 pst 2001
63. e yet powerful XML markup tag called lt srai gt Symbolic Reduction Artificial Intelligence Syntactic Rewrite Al Simple Recursive Al Stimulus Response Al this tag has many acronyms Yet the meaning of the markup lt srai gt X lt srai gt is simple The lt srai gt tag always appears in the response template but the robot treats X just like an input to the robot The robot scans through its memory and finds the best response for X The only tricky part is the response to Xmay itself contain more lt srai gt tags The best way to understand the recursive action of the AIML lt srai gt tag is by example Client You may say that again Alice Robot Once more that The robot has no specific response to the pattern You may say that again Alice Instead the robot builds its response to the client input in four steps This simple sentence activated a sequence of four categories linked by lt srai gt tags The robot constructed the reply Once more that recursively as each subsentence triggered the next matching pattern In this example the processing proceeds in four steps because each of the first three steps evokes another symbolic reduction Step normalized input matching pattern template response 1 YOU MAY SAY THAT AGAIN ALICE _ name sr 2 YOU MAY SAY THAT AGAIN AGAIN Once more sr Once More 3 YOU MAY SAY THAT YOU MAY lt sr gt Once More 4 SAY THAT SAY lt person gt Once More
64. ecreases with each successive word 5831 WHAT IS YOUR NAME 352 WHAT IS MY NAME 171 WHAT IS UP 137 WHAT IS YOUR FAVORITE COLOR 126 WHAT IS THE MEANING OF LIFE 122 WHAT IS THAT 102 WHAT IS YOUR FAVORITE MOVIE 92 WHAT IS IT 75 WHAT IS A BOTMASTER 70 WHAT IS YOUR IQ 59 WHAT IS REDUCTIONISM 53 WHAT IS YOUR FAVORITE FOOD 46 WHAT IS AIML 38 WHAT IS YOUR FAVORITE BOOK 37 WHAT IS THE TIME 37 WHAT IS YOUR JOB 34 WHAT IS YOUR FAVORITE SONG 34 WHAT IS YOUR SIGN 33 WHAT IS SEX 32 WHAT IS YOUR REAL NAME 30 WHAT IS NEW 30 WHAT IS YOUR AGE 30 WHAT IS YOUR GENDER 28 WHAT IS YOUR LAST NAME 27 WHAT IS HIS NAME 27 WHAT IS YOUR SEX 26 WHAT 15 2 2 26 WHAT IS MY IP 25 WHAT IS YOURS 24 WHAT IS YOUR PURPOSE 21 WHAT IS YOUR FAVORITE ANIMAL 20 WHAT IS 1 1 20 WHAT IS YOUR HOBBY 19 WHAT IS THE WEATHER LIKE 19 WHAT IS YOUR PHONE NUMBER 18 WHAT IS ALICE 18 WHAT IS GOING ON 18 WHAT IS THAT SUPPOSED TO MEAN 18 WHAT IS WHAT 17 WHAT IS A SEEKER 17 WHAT IS LOVE 17 WHAT IS THE OPEN DIRECTORY 17 WHAT IS YOUR FAVORITE TV SHOW 16 WHAT IS JAVA 16 WHAT IS THE ANSWER 16 WHAT IS YOUR ANSWER 16 WHAT IS YOUR FULL NAME 15 WHAT IS AI 15 WHAT IS THAT MEAN 15 WHAT IS THE WEATHER LIKE WHERE YOU ARE 15 WHAT IS TWO PLUS TWO 15 WHAT IS YOUR FAVORITE BAND 14 WHAT IS CBR 14 WHAT IS ELIZA 14 WHAT IS GOD 14 WHAT 15 PI 14 WHAT IS THE TURING GAME 13 WHAT 15 2 2 13 WHAT IS A COMPUTER YEAR 13 WHAT IS IT LIKE 13 WHAT IS MY FA
65. ething unexpected or puts together responses in ways the botmaster never indented For example Client bet you are gay ALICE Actually am not the gambling type Actually as a machine have no need for sex Here the robot linked two different categories which both coincidentally have a moral theme gambling and sexuality But this specific combination was not preprogrammed by the botmaster Are these surprising responses just unintended coincidences or do they indicate that ALICE is thinking Is ALICE just a gigantic stimulus response mechanism or are we What is the theory behind ALICE used to say that there was NO theory behind ALICE no neural network no knowledge representation no search no fuzzy logic no genetic algorithms and no parsing Then discovered there was a theory circulating in applied called Case Based Reasoning or CBR that maps well onto the ALICE algorithm Another term borrowed from pattern recognition is nearest neighbor classification The CBR cases are the categories in AIML The algorithm finds best matching pattern for each input The category ties the response template directly to the stimulus pattern ALICE is conceptually not much more complicated that Weizenbaum s ELIZA chat robot the main differences are the much larger case base and the tools for creating new content by dialog analysis ALICE is also part of the tradition of minimalist reactive or stimulus respons
66. fer them directly to your checking account This process may take a few days depending on your bank PayPal has several other options for withdrawing funds check their website On Pandorabots go to the Botmaster Control page In the column marked Subscribers click on the Subscribers link On the Subscribers page you will see a form with two text fields to Add a New 79 Subscriber Enter the email address and the number of months this customer will subscribe After you click the Add subscriber button Pandorabots will update the table to display a new row corresponding to the new customer s bot link Under the column marked Access URL click on Subscriber s bot and follow the link The URL that appears in the browser is the customer s private subscription link This is the link that you will email to the customer and tell them not to share with anyone else After you have collected the payment and created the subscriber s bot it is time to send them their private link Cut and paste the URL from the browser see previous step Send the customer the link in an email message It is a good idea to have a form letter ready like this one explaining to the customer that they should bookmark the link and keep it private If they lose it they can contact you and you can resupply it What about Refunds and Deactivation In any business some customers are bound to come back dissatisfied with their purchases You should be prepare
67. get name he gt lt template gt lt category gt category gt pattern WHOISHE UNKNOWN lt pattern gt template don t know who is lt template gt lt category gt Provi ded that the predicate he is initialized to Unknown the categories execute a conditional branch depending on whether he has been set As a convenience to the botmaster AIML also provides the equivalent function through the lt condition gt tag TARGETING Bot personalities are created and shaped through a cyclical process of supervised learning called Targeting Targeting is a cycle incorporating client bot and botmaster wherein client inputs that find no complete match among the categories are logged by the bot and delivered as Targets the botmaster who then creates suitable responses starting with the most common queries The Targeting cycle produces a progressively more refined bot personality The art of AIML writing is most apparent in creating default categories which provide noncommittal replies to a wide range of inputs Here is a very nice example provided found on the placebo general mailing list User Where do find the web page with Australopithecus 16 Alice Where do you want to find it User Bad Answer Alice time to Learn a new Line Alice How would you answer Where do l find the web page with Australopithecus User would say http www ancientskulls net australopithecus html Alice Thanks hum
68. gets the botmaster sees Input WHAT IS PIZZA OH READY Matched WHAT IS New AIML WHAT IS PIZZA The targeting algorithm extended the existing pattern by one word from the input to obtain the new AIML pattern WHAT IS PIZZA In the template window the botmaster types Pizza is a delicious food made with dough cheese and tomato sauce The botmaster may also click the think button to add the text delicious food made with dough cheese and tomato sauce think set name it gt set name topic gt 42 person lt set gt lt set gt lt think gt But s he needs to edit the text to achieve the desired result of remembering the topic is PIZZA Pizza is a delicious food made with dough cheese and tomato sauce lt think gt set gt lt set name topic gt PIZZA lt set gt lt set gt lt think gt After entering this text in the lt template gt text area the botmaster may save the new category with the Save Category button 2 In the second example the client entered WHAT IS A PIZZA the topic is PIZZA and the value of lt that gt is HAVE TO PROCESS THAT ONE FOR A WHILE The Targeting display shows Input WHAT IS A PIZZA HAVE TO PROCESS THAT ONE FOR A WHILE PIZZA Matched WHAT IS A New AIML WHAT IS A PIZZA Here the targeting algorithm again extended the existing pattern by one word taken from the input creating the
69. gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category pattern l PUT DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt l CHARGED DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 3 gt lt srai gt lt template gt
70. gt lt category gt category gt lt pattern gt I PAID FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON AT lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt
71. gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED ON MY ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category pattern l CHARGED ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category pattern l PUT DOLLARS ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category pattern l PU
72. gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY ATM FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star 73 index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT MY ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ATM CARD ON FOR lt pattern gt lt template
73. hart but overall the phone curve is more like a long straight line than an S curve Cable TV is also flatter than many others From its inception in 1976 you can almost draw a straight line to its present position at 6896 market share The PC also runs in a straight line from the coordinates 1980 0 to 2000 45 In some sense this is why the Cable TV and PC businesses are so predictable and boring Winning over that last remaining half of households is just a matter of time and not making any big mistakes It also tells you why people like Bill Gates always want to expand into new business areas like games and entertainment Even when the PC market saturates Bill can only be twice as rich as he is now To get more he needs another line of business What does all of this have to do with L I C E and AIML People often imagined that A L I C E is like Netscape in 1995 poised to take off on the next great S curve of growth But have argued it is more like Apple in 1975 or the whole PC industry at that time We are mainly amateurs and hobbyists with few real business applications we can point to Yet we all have the same kind of fever that gripped an earlier generation of PC enthusiasts We KNOW this is the Next Big Thing We KNOW we are the vanguard of the revolution The first Apple appeared in 1975 But the technology adoption curve for PCs doesn t even begin until 1980 Those early PC pioneers were like Ph
74. his example borrowed from ELIZA category lt pattern gt MOTHER lt pattern gt lt template gt Tell me more about your family lt template gt lt category gt lt category gt patterns MOTHER lt pattern gt lt template gt lt srai gt MOTHER lt srai gt lt template gt lt category gt lt category gt lt pattern gt MOTHER _ lt pattern gt lt template gt lt srai gt MOTHER lt srai gt lt template gt lt category gt lt category gt lt pattern gt _ MOTHER lt pattern gt lt template gt lt srai gt MOTHER lt srai gt lt template gt 15 lt category gt The first category both detects the keyword when it appears by itself and provides the generic response The second category detects the keyword as the suffix of a sentence The third detects it as the prefix of an input sentence and finally the last category detects the keyword as an infix Each of the last three categories uses lt srai gt to link to the first so that all four cases produce the same reply but it needs to be written and stored only once CONDITIONALS It is possible to write conditional branches in AIML using only the lt srai gt tag Consider three categories lt category gt lt pattern gt WHO IS HE lt pattern gt lt template gt lt srai gt WHOISHE get name he gt lt srai gt lt template gt lt category gt category gt lt pattern gt WHOISHE lt pattern gt lt template gt He is
75. i gt lt think gt lt gossip gt DEBIT lt star gt lt star index 2 gt lt star index 3 gt lt star index 4 gt star index 5 gt star indexz 6 star index 7 gt lt id gt lt gossip gt lt think gt lt template gt lt category gt lt category gt lt pattern gt INITIALIZE BANK ACCT AMT lt pattern gt lt template gt srai ACKNOWLEDGE TRANSACTION lt srai gt lt think gt lt gossip gt INITIALIZE BANK lt star gt lt star index 2 gt lt date gt lt id gt lt gossip gt lt srai gt CREDIT ACCT lt star gt AMT lt star index 2 gt DATE lt date gt lt srai gt lt think gt lt template gt lt category gt 61 category gt lt pattern gt INITIALIZE CREDT ACCT AMT pattern template srai ACKNOWLEDGE TRANSACTION lt srai gt think lt gossip gt INITIALIZE CREDT star star index 2 gt lt date gt lt id gt lt gossip gt srai DEBIT ACCT lt star gt AMT star index 2 gt DATE lt date gt lt srai gt lt think gt lt template gt lt category gt category gt lt pattern gt BALANCE ACCT lt pattern gt lt template gt lt system gt sh home alicebot balance sh lt star gt lt system gt lt template gt lt category gt lt category gt lt pattern gt HOW MUCH IS IN MY WALLET lt pattern gt lt template gt lt srai gt BALANCE ACCT WALLET srai template category gt category gt lt g
76. ilo Farnsworth and his TV picture tube They knew that they had a great thing but getting people to adopt it was like feeding them liver Jobs Wozniak Gates and friends only had to wait five years compared to Farnsworth s twenty l ve been at A L I C E and AIML for almost seven years now don t know if we are on the knee on the ramp or if the adoption curve has even started yet have learned in seven years not to expect miracles and to have patience Unfortunately we need the perspective of ten or twenty years to see what is really happening here conclude with a quote from Kawasaki s book Don t worry about people stealing your ideas If your ideas are any good you ll have to ram them down people s throats Howard Aiken REFERENCES Prolog References 1 Prolog Tutorial http cs wwc edu cs dept KU PR Prolog html 83 2 A Short Tutorial on Prolog http cbl leeds ac uk tasmin prologtutorial 3 Prolog Programming A First Course http cbl leeds ac uk paul prologbook 4 GNU Prolog site http pauillac inria fr diaz gnu prolog References on Zipf s Law http www cut the knot com do_you_know zipfLaw html http linkage rockefeller edu wili zipf http www useit com alertbox zipf htm 1 Documentation http alicebot org documentation 2 Don t Read Me Program dB http alicebot org articles wallace dont dB html 3 AIML Specification Working Draft http alicebot org TR 2001 WD aiml ACK
77. ing AMT star srai template category gt category pattern DEPOSIT TO CHECKING lt pattern gt lt template gt lt srai gt CREDIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category pattern DEPOSIT IN MY SAVINGS ACCOUNT x pattern template srai CREDIT ACCT savings AMT star srai template category category pattern DEPOSIT SAVINGS ACCOUNT x pattern template srai CREDIT ACCT savings AMT star srai template category gt category pattern DEPOSIT IN SAVINGS pattern template srai CREDIT ACCT savings AMT lt star gt lt srai gt lt template gt lt category gt category pattern DEPOSIT TO SAVINGS ACCOUNT x pattern template srai CREDIT ACCT savings AMT star srai template category gt category pattern DEPOSIT TO SAVINGS pattern template srai CREDIT ACCT savings AMT lt star gt lt srai gt lt template gt lt category gt 64 category pattern CREDIT ACCT AMT lt pattern gt lt template gt lt srai gt CREDIT ACCT lt star gt AMT star index 2 gt DATE lt date gt lt srai gt lt template gt lt category gt category pattern CREDIT ACCT AMT DATE lt pattern gt lt template gt lt srai gt CREDIT ACCT lt star gt AMT star index 2 gt DATE star index 3 gt NUMBER NONE lt srai gt lt template gt
78. lance is 2300 Robot got it Do you have any more transactions Client My savings account balance is 1000 Robot got it Anything else can do for you Client My visa card balance is 2100 Robot got it Thanks for the information Client My Citibank balance is 9000 Robot got it Is that it Client have 200 dollars in my wallet Robot Alright Thanks for the information Client spent 10 on a taxi Robot got it Anything else can do for you Client charged 23 for lunch on my visa card Robot got it Thanks for the information Client My landlord is Mr bill Robot Rent is theft Client No kidding my rent is 1410 a month 58 Robot Would kid you 1410 is the market rate Client paid the rent with check number 101 Robot got it Thanks for the information Client what is my checking account balance Robot Your checking balance is 890 Client charged 120 on my Visa card for the hotel Robot Ok Anything else can do for you Client What is my Visa card balance Robot Your visa balance is 2243 Client How much money do have in my wallet Robot Your wallet balance is 190 Client bought a book for 10 Robot Yes have recorded your transaction Client now how much is in my wallet Robot Why now Your wallet balance is 180 THE TRANSACTIONS The preceding dialogue reduces to the transactions listed in this section The DEBIT and CREDIT transactions have exac
79. late gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY ATM CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED ON MY AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON MY ATM CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS MY ATM CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NU
80. ledge are stated in the simplest possible terms for example we tend to prefer patterns like WHO IS SOCRATES 13 to ones like DO YOU KNOW WHO SOCRATES IS when storing biographical information about Socrates Many of the more complex forms reduce to simpler forms using AIML categories designed for symbolic reduction category lt pattern gt DO YOU KNOW WHO IS pattern template srai WHO IS lt star gt lt srai gt lt template gt lt category gt Whatever input matched this pattern the portion bound to the wildcard may be inserted into the reply with the markup lt star gt This category reduces any input of the form Do you know who X is to Who is X DIVIDE AND CONQUER Many individual sentences may be reduced to two or more sub sentences and the reply formed by combining the replies to each A sentence beginning with the word Yes for example if it has more than one word may be treated as the sub sentence Yes plus whatever follows it lt category gt lt pattern gt YES lt pattern gt lt template gt lt srai gt YES lt srai gt lt sr gt lt template gt lt category gt The markup lt sr gt is simply an abbreviation for lt srai gt lt star gt lt srai gt SYNONYMS The AIML 1 01 standard does not permit more than one pattern per category Synonyms are perhaps the most common application of lt srai gt Many ways to say the same thing reduce to one category which co
81. lly matter which personality classification system you pick In some sense what you said is true they are all bunk Or maybe 81 they seem to work 60 of the time for 6096 of the people perhaps Now suppose the first botmaster menu is Do you prefer Luscher Color Quiz Astrology Enneagram Meyers Briggs Pick a celebrity Homeopathy None of the above The Personality Test for the Astrology choice is simple What is your birthday Now it doesn t matter if the Virgo bot personality really matches the botmaster s personality or not as long as the botmaster believes it does The customer is always right If the Four botmaster believed the Enneagram system was more reliable then she might get a better result with the Enneagram Four bot If the botmaster believed that all personality classifications are bunk then they could do it the old fashioned way we do it now with None of the Above CONCLUDING REMARKS have a very interesting chart of technology adoption during the last century It was originally produced by the Wall Street Journal but 1 have a small copy reproduced in the book Rules for Revolutionaries by Guy Kawasaki Harper 1999 The chart shows for a variety of technologies from the telephone to satellite TV the growth of each in the marketplace as a function of year The Y axis is expressed in percentage of U S households so none of the curves exceeds 100 Generally all of the curves follow
82. lt gt SPENT DOLLARS MY ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE date srai template category gt category gt lt pattern gt I SPENT ON MY CARDz pattern template srai DEBIT ACCT checking AMT lt gt DATE date srai template category gt category gt lt pattern gt I SPENT ON MY ATM x pattern template srai DEBIT ACCT checking AMT star DATE date srai template category gt category gt lt pattern gt I PAID DOLLARS WITH CHECK NUMBER pattern template srai DEBIT ACCT checking AMT star index 2 gt DATE lt date gt NUMBER star index 3 gt PAYEE lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt l WROTE CHECK NUMBER FOR TO pattern template srai DEBIT ACCT checking AMT star index 2 gt DATE lt date gt NUMBER lt star gt PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt gt WROTE A CHECK TO FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star index 2 gt DATE lt date gt NUMBER UNKNOWN PAYEE lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt l WROTE CHECK NUMBER TO FOR pattern template srai DEBIT ACCT checking AMT star index 3 gt DATE lt d
83. lt category gt category pattern CREDIT ACCT AMT DATE NUMBER PAYER CATEGORY pattern template srai CREDIT ACCT lt star gt AMT star index 2 gt DATE star index 3 gt NUMBER star index 4 gt PAYER star index 5 gt CATEGORY star index 6 gt NOTE NONE lt srai gt lt template gt lt category gt category pattern CREDIT ACCT AMT DATE NUMBER PAYER pattern template srai CREDIT ACCT lt star gt AMT star index 2 gt DATE star index 3 gt NUMBER star index 4 gt PAYER star index 5 gt CATEGORY UNKNOWN lt srai gt lt template gt lt category gt category pattern CREDIT ACCT AMT DATE NUMBER pattern template srai CREDIT ACCT lt star gt AMT star index 2 gt DATE star index 3 gt NUMBER star index 4 gt PAYER UNKNOWN lt srai gt lt template gt lt category gt category gt lt pattern gt l DEPOSITED IN lt pattern gt lt template gt lt srai gt CREDIT ACCT star index 2 gt AMT lt star gt srai template category gt category gt lt pattern gt l DEPOSITED INTO lt pattern gt lt template gt lt srai gt CREDIT ACCT star index 2 gt AMT lt star gt srai template category gt category gt lt pattern gt l DEPOSITED MY pattern template srai CREDIT ACCT star index 2 gt AMT star srai template category gt ca
84. lt category gt category gt lt pattern gt I SPENT ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT star DATE lt date gt lt srai gt lt template gt lt category gt lt category gt lt pattern gt I CHARGED MY CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt lt srai gt lt template gt lt category gt 71 category gt lt pattern gt I PUT DOLLARS ON MY CARDx pattern template srai DEBIT ACCT star index 2 gt AMT star DATE lt date gt lt srai gt lt template gt lt category gt category pattern l PUT ON MY CARD pattern2 template srai DEBIT ACCT star index 2 gt AMT star DATE lt date gt lt srai gt lt template gt lt category gt category pattern SPENT DOLLARS ON MY CARD pattern template srai DEBIT ACCT star index 2 gt AMT star DATE lt date gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD pattern template srai DEBIT ACCT star index 2 gt AMT st
85. lt that gt encloses a pattern that matches the robot s previous utterance A common application of that is found in yes no questions category pattern YES pattern 17 that DO YOU LIKE MOVIES lt that gt template What is your favorite movie lt template gt lt category gt This category is activated when the client says YES The robot must find out what is he saying yes to If the robot asked Do you like movies this category matches and the response What is your favorite movie continues the conversation along the same lines One interesting application of that are categories that enable the robot to respond to knock knock jokes The categories category pattern KNOCK KNOCK lt pattern gt template Who is there lt template gt lt category gt category gt lt pattern gt lt pattern gt lt that gt WHO IS THERE lt that gt lt template gt lt person gt who lt template gt lt category gt lt category gt lt pattern gt lt pattern gt lt that gt WHO lt that gt lt template gt Ha ha very funny lt get name name gt lt template gt lt category gt produce the following dialogue Knock knock Who s there Banana Banana who Knock knock Who s there Banana Banana who Knock knock Who s there Orange Orange who D O DO D D O P Orange you glad didn t say banana 18
86. lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED DOLLARS ON MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON CARD FOR 67 lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category
87. ly in the hinterlands occupied by only the rarest forms of language Our territory of language already contains the highest population of sentences that people use Expanding the borders even more we will continue to absorb the stragglers outside until the very last human critic cannot think of one sentence to fool A L I C E 20 CHAPTER Il AIML Pattern Matching This chapter presents an of explanation of how the Graphmaster works for those who want to build their own or simply want a deeper understanding of the approach THE GRAPHMASTER The Graphmaster consists of a collection of nodes called Nodemappers These Nodemappers map the branches from each node The branches are either single words or wildcards The root of the Graphmaster is a Nodemapper with about 2000 branches one for each of the first words of all the patterns 40 000 in the case of the A L I C E brain The number of leaf nodes in the graph is equal to the number of categories and each leaf node contains the template tag There are really only three steps to matching an input to a pattern If you are given a an input starting with word X and b a Nodemapper of the graph Does the Nodemapper contain the key If so search the sub graph rooted at the child node linked by Try all remaining suffixes of the input following X to see if one matches If no match was found try Does the Nodemapper contain the key X If so search the
88. nce the security feature to prohibit unauthorized clients from making personal financial transactions f Substitute voice for text input This is the kind of application you want to carry around with you so you can utter I spent 10 on a taxi at the same time you pay the driver and l took out 100 from the ATM while you are still at the cash machine More generally we need to work AIML APIs for a variety of everyday applications These include but are not limited to address books spreadsheets databases email web browsers word processors and games In the dream world of the future the talking Star Trek HAL style computer will handle all of the applications we use now on WIMP Windows Icons Menus Pointing Device computers Appendix A AIML categories for personal finance Note For the personal finance application these categories should be included along with those listed in section 4 Most of the categories have been moved to an appendix so that this document is not too long category pattern DEPOSIT CHECKING pattern template srai CREDIT ACCT checking AMT star srai template category gt category pattern DEPOSIT MY CHECKING ACCOUNT x pattern template srai CREDIT ACCT checking AMT star srai template category gt category pattern DEPOSIT TO CHECKING ACCOUNT pattern template srai CREDIT ACCT check
89. ng AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I TOOK OUT AT THE ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt I GOT OUT AT THE ATM lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt lt srai gt template category gt category pattern l CHARGED DOLLARS ON MY CARD ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED MY CARD ON FOR pattern template srai DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt PUT DOLLARS ON MY CARD ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY CARD ON
90. ng engine Everything described here took place on a Linux Red Hat 7 1 machine We have not yet tried running this under Windows although there is a Windows version of gprolog available The goal was to create a set of AIML categories that answer questions using a Prolog program This experiment implements a simple isa hierarchy so the bot can answer questions like Is duck a food Is rock alive Is a dog an animal Is a cat a pet and so on The system command The basic approach was to interface AIML to Prolog through the AIML system tag The Prolog program ran as a separate executable activated by the system tag It was not too difficult to download and install the GNU gprolog interpreter and compiler from http pauillac inria fr diaz gnu prolog There are HTML documentation pages and several directories containing programming examples The associated Makefiles were a useful guide to writing and compiling simple Prolog programs The gprolog suite includes a compiler to create executable prolog programs The PSYCH script was compiled into an executable called psych Normally Prolog runs in an interactive prompting mode but it is possible to redirect the input from a file or with a pipe 45 The basic form of a prolog query is like isa dog animal or more generally predicate term1 term2 The ending period is important playing much the same role as the ending in
91. ntains the reply lt category gt lt pattern gt HELLO lt pattern gt lt template gt Hi there lt template gt lt category gt lt category gt lt pattern gt Hl lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt lt category gt lt pattern gt HI THERE lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt 14 category gt patterns HOWDY pattern template srai HELLO srai template lt category gt lt category gt lt pattern gt HOLA lt pattern gt lt template gt lt srai gt HELLO lt srai gt lt template gt lt category gt SPELLING AND GRAMMER CORRECTION The single most common client spelling mistake is the use of your when you re or you are is intended Not every occurrence of your however should be turned into you re A small amount of grammatical context is usually necessary to catch this error lt category gt lt pattern gt YOUR A lt pattern gt lt template gt think you mean you re or you are not your srai YOU ARE A lt star gt lt srai gt template category gt Here the bot both corrects the client input and acts as a language tutor KEYWORDS Frequently we would like to write an AIML template that is activated by the appearance of a keyword anywhere in the input sentence The general format of four AIML categories is illustrated by t
92. o ALICE we need to visit another unusual figure in the history of computer science Professor George Kingsley Zipf Although he was a contemporary of Turing there is no evidence the two ever met Zipf died young too at the age of 48 in 1950 only four years before Turing but of natural causes There are many ways to state Zipf s Law but the simplest is procedural Take all the words in a body of text for example today s issue of the New York Times and count the number of times each word appears If the resulting histogram is sorted by rank with the most frequently appearing word first and so on a the for by and then the shape of the curve is Zipf curve for that text If the Zipf curve is plotted on a log log scale it appears as a straight line with a slope of 1 The Zipf curve is a characteristic of human languages and many other natural and human phenomena as well Zipf noticed that the populations of cities followed a similar distribution There are a few very large cities a larger number of medium sized ones and a large number of small cities If the cities or the words of natural language were randomly distributed then the Zipf curve would be a flat horizontal line The Zipf curve was even known in the 19th century The economist Pareto also noticed the log rank property in studies of corporate wealth One only needs to consider the distribution of wealth among present day computer companies to see th
93. of art lt template gt lt category gt lt category gt pattern SINGULAR WRENCHES lt pattern gt lt template gt wrench lt template gt lt category gt The singular function also requires a default case when the word or words are not recognized in the function domain lt category gt pattern SINGULAR lt pattern gt template star lt template gt lt category gt Notice that if Singular X Y is not defined i e this last category matches then Isb X Y Isa X Y RANDOM INFERENCES The categories described in sections 1 5 provide answers to questions like WHAT DOES X HAVE by first checking to see if there is an explicit match for WHAT DOES X HAVE If not the program proceed to WHAT DOES Y HAVE where Isa X Y If no match is found there the program proceeds to WHAT DOES Z HAVE where Isa Y Z Because there is an ultimate WHAT DOES UNKNOWN HAVE category the procedure will always terminate The inferences will not however find ALL instances of the logical reply to WHAT DOES X HAVE For example if we know that all animals have a mother then one answer to What does a bird have should logically be a mother The following two categories show a crude but simple way to overcome this obstacle If the client asks What else 56 do birds have the random function will select between two possible alternative questions What do birds have and for some X such that Isa bird X
94. pth first search also known as backtracking You can simplify the algorithm by removing the wildcard and considering just the second two steps Also try understanding the simple case of Patterns without that and topic The File system Metaphor A convenient metaphor for AIML patterns and perhaps also an alternative to database storage of patterns and templates is the file system Hopefully by now almost everyone understands that their files and folders are organized hierarchically in a tree Whether you use Windows Unix or Mac the same principle holds true The file system has a root such as c Y The root has some branches that are files and some that are folders The folders in turn have branches that are both folders and files The leaf nodes of the whole tree structure are files Some file systems have symbolic links or shortcuts that all you to place virtual backward links in the tree and turn it into a directed graph but forget about that complexity for now Every file has a path name that spells out its exact position within the tree c my documents pictures me jpg denotes a file located down a specific set of branches from the root The Graphmaster is organized in exactly the same way You can write a pattern like I LIKE TO as g I LIKE TO star All of the other patterns that begin with I also go into the g I folder of the patterns that begin with I LIKE go in the g I LIKE
95. pting credit card or bank account payments for purchases If you have an email account you can create an account on PayPal with the same name For example if your email name is artjohnson xyz ww then your PayPal business name would also be artjohnson xyz ww You can then invite people to send you money from their PayPal accounts or send money to their accounts Go to www PayPal com and click on the link that says Sign up for your free PayPal account You have to fill out a one page form of information You only need your email address to create an account Later you can add your checking account and or credit cards The PayPal system will email you a verification letter After you complete the verification process your PayPal account will be activated and you can send and receive funds electronically Pandorabots http www pandorabots com is a software robot also known as a bot hosting service From any browser you may create and publish your own robots to anyone via the web We believe that our technology yields the fastest bots available on the Internet The bots are based on AIML and spring entirely from the work of Dr Richard Wallace and the A L C E and free software community based at http www alicebot org There is no charge for using this site We gladly accept any inquiries for custom versions of this site for 78 particular applications Send mail to info pandorabots com with questions No one has
96. r hand is designed to keep the client talking as long as possible without necessarily providing any useful information along the way The longer average conversation lengths measured over the years have in fact been a measure of A L I C E s progress Strange as it seems free software is still not considered a mainstream methodology for either research or software development As big and successful as Linux is the majority of the world has not yet bought the free software argument For many people the whole debate over software intellectual property remains obscure Add artificial intelligence to the mix and you have something truly unique There are a few other free software Al projects out there but none has had the impact of Alicebot AIML This year was our first appearance at Linuxworld a first step in finding common ground between the Al and free software communities Finally many in the Alicebot AIML community share a common vision of the future of our technology Alicebot is perceived as a missing piece of the puzzle that combines speech recognition natural language understanding and voice synthesis The vision is the talking Star Trek HAL style computer of the future the invisible piece of hardware that responds to voice commands In that vision there is no keyboard display mouse or graphical user interface If that vision comes true the impact of AIML will be ubiquitous WHAT MAKES ALICE WORK Before we get t
97. r or other toy Falling prices of embedded micro controllers and DSPs make toys practical 4 Celebrity Bot Fans of Britney Spears can go to her fan site and chat with Britney Bot ELVIS and John Lennon have already been done and there is likely to be 77 a land rush on deceased celebrity names Timothy Leary Andy Warhol Ronald Reagan JFK Frank Zappa Richard Nixon Malcom X Martin Luther King Jr John Travolta O J Simpson Albert Einstein Marilyn Monroe Picasso Janis Joplin and finitely many more 5 Personal Assistant Voice based personal secretary handles finances address book database and email All of the most common computer applications will have natural language voice based interfaces 6 IM Plugin Instant Messaging users can set their chat connection to automatic or manual 7 Targeting Tool Programs to assist the botmaster analyze the log files and add bot content 8 Bot Hosting Many people cannot afford 24 7 Web connections but would like to have their bots always online 9 Social Security Bot Many disabled and retiring people have basic questions about benefits applications eligibility and other aspects of the Social Security process 10 Go Ask ALICE Harm Reduction Bot Help reduce the harmful effects of hard drug use by providing truthful information to drug users PayPal allows businesses and consumers with e mail addresses to send and receive payments via the Internet acce
98. rain isa muni train isa tram train isa bicycle vehicle isa bus vehicle isa truck vehicle isa suv truck isa rocket ehicle isa vehicle manmade isa manmade entity isa puddle water isa ocean water isa lake water isa river water isa stream water va ava mcam m o mo mn 51 isa sea water isa cloud water sa water compound a compound entity sa ice water sa steam water isa aspirin medicine o sa isa penicillin medicine sa cancer disease sa aids disease isa flu disease isa glaucoma disease CHAPTER VI Doing Simple Logical Deductions in AIML The purpose of this exercise is to demonstrate some simple reasoning capabilities with AIML alone In particular we did not use any logic engine such as Prolog The results described here were obtained solely with AIML The types of questions under consideration here are a restricted set of queries exemplified by What does a bird have What do birds have What does a raven do What else does a bird have What else do ravens do Part of what makes it easy to answer these questions in AIML is that the questions tend to have multiple answers A bird has a beak a tail lungs eyes wings feathers and cold blood It also has all the properties of any superclasses such as animals or prokaryotes The AIML program does not need to find all of these solutions
99. random gt lt template gt lt category gt REDUCTIONS In this step we apply symbolic reductions to transform a variety of grammatical forms into simpler inputs lt category gt lt pattern gt WHAT DO DO lt pattern gt lt template gt lt srai gt WHAT DOES A DO lt srai gt lt template gt lt category gt lt category gt lt pattern gt WHAT DOES A DO lt pattern gt lt template gt lt srai gt WHAT DOES lt star gt DO lt srai gt 54 lt template gt lt category gt These reductions along with others like them will transform many grammatical variants like WHAT DOES A X DO WHAT DOES AN X DO and WHAT DO X DO into a single canonical form like WHAT DOES X DEDUCTIONS Provided that the AIML interpreter is capable of executing nested srai tags it is quite possible to make simple inferences in AIML The logic presented here is that if the client asks WHAT DOES X DO and there is no specific answer for X then the program will try to answer the question WHAT DOES Y DO where Isa X Y The corresponding logical deduction would be If X is a Y and Y does Z then X does Z The deduction may be written in AIML as category pattern WHAT DOES DO lt pattern gt template srai WHAT DOES lt srai gt ISA star srai DO lt srai gt template lt category gt The similar category for WHAT DOES X HAVE is lt category gt lt p
100. s any other for these cases and certainly the simplest If there is any room left for higher natural language theories it lies outside the map of the ALICE brain Academics are fond of concocting riddles and linguistic paradoxes that supposedly show how difficult the natural language problem is John saw the mountains flying over Zurich or Fruit flies like a banana reveal the ambiguity of language and the limits of an ALICE style approach though not these particular examples of course ALICE already knows about them In the years to come we will only advance the frontier further The basic outline of the spiral graph may look much the same for we have found all of the big trees from A to YOUR These trees may become bigger but unless language itself changes we won t find any more big trees except of course in foreign languages The work of those seeking to explain natural language in terms of something more complex than stimulus response will take place beyond our frontier increasingly in the hinterlands occupied by only the rarest forms of language Our territory of language already contains the highest population of sentences that people use Expanding the borders even more we will continue to absorb the stragglers outside until the very last human critic cannot think of one sentence to fool ALICE 32 CHAPTER Symbolic Reductions in AIML AIML Artificial Intelligence Markup Language contains a simpl
101. s do Robot Peck A bird has a nuclear membrane because it is a prokaryote All animals have a body so birds have a body too Chicken is both 57 a food and a bird so it can both cluck and be eaten Ravens can fly because the robot believes all birds can fly LOGICAL DEDUCTIONS IN AIML We have shown how to make certain simple logical deductions in AIML alone without the need for an external Prolog or other reasoning program The ALICE brain set was augmented with more than 1000 Isa relations and several hundred categories encoding Has and Does These categories will be made public in the next release of the ALICE Brain AIML set CHAPTER VII A Personal Finance Spreadsheet in AIML One of the critical requirements for the adoption of AIML technology as a standardized human computer interface is the creation of a core set of application programming interfaces For ALICE to become the primary mode of user interaction she must have interfaces to a collection of common applications such as email database address book calculator and spreadsheet programs This chapter describes a first pass at the spreadsheet application for managing personal finances SAMPLE DIALOG The goal of this experiment is to produce a believable natural language interface to a trivial personal finance System The following sample dialogue illustrates the flavor of conversations the bot should be able to handle Client My checking account ba
102. s is non empty For this purpose we use the Prolog built in predicate setof isaset X Y setof isb X Y The query based on the predicate isaset is called 47 by the shell script described above The isaset predicate returns Yes or No depending on whether a set of solutions of isb X Y exists THE AIML CATEGORIES The AIML category used to test the Prolog program PSYCH is category lt pattern gt IS A A lt pattern gt template system sh home alicebot isa sh star index 1 gt star index 2 gt lt system gt lt template gt lt category gt Note that rather than just run the script directly Java Runtime exec forced us to prepend a sh command to the line The values of lt star index 1 gt and lt star index 2 gt become the arguments to isa sh Next a few lt srai gt categories map variations of the logical isa question to the simple pattern IS A A category lt gt 15 A AN lt pattern gt lt template gt lt srai gt IS lt star gt A star index 2 gt lt srai gt lt template gt category gt Similar lt srai gt categories were defined for lt pattern gt IS AN A lt pattern gt lt pattern gt IS AN lt pattern gt lt pattern gt IS lt pattern gt lt pattern gt IS AN lt pattern gt and lt pattern gt IS lt pattern gt SAMPLE DIALOG The following is a sample dialogue be
103. s loaded The spine is actually a log Spiral but with an exponent close to unity so very nearly the same as a linear spiral Gray lines indicate nodes with exactly one branch Black lines are nodes with two or more branches The leaf nodes have two branches because they store both template and filename 30 This squared spiral was formed by modulating the log spiral with a sine wave The resulting figures uses more of the display area than the raw log spiral The plot also shows the words sorted by word length which results in a nice balance among the subtree graphs Large subtrees tend to begin with shorter words like A I YOU and WHAT Longer words tend to have fewer branches Arranging the graph in this way the branches near the center of the spiral are generally fewer than the branches near the outside The picture makes it easier to see the subtrees than say an alphabetical ordering which agglomerates many subtrees at the center More than just an elegant graph of the ALICE brain these spiral images outline a territory of language that has been effectively conquered by A L C E and AIML No other theory of natural language processing can better explain or reproduce the results within our territory You don t need a complex theory of learning neural 3l nets or cognitive models to explain how to chat within the limits of ALICE s 40 000 categories Our stimulus response model is as good a theory a
104. s simple as possible for the task at hand simpler even than regular expressions The template is the AIML response or reply In its simplest form the template consists of only plain unmarked text More generally AIML tags transform the reply into a mini computer program which can save data activate other programs give conditional responses and recursively call the pattern matcher to insert the responses from other categories Most AIML tags in fact belong to this template side sub language AIML currently supports two ways to interface other languages and systems The system tag executes any program accessible as an operating system shell command and inserts the results in the reply Similarly the javascript tag allows arbitrary scripting inside the 12 templates The optional context portion of the category consists of two variants called that and topic The that tag appears inside the category and its pattern must match the robot s last utterance Remembering one last utterance is important if the robot asks a question The topic tag appears outside the category and collects a group of categories together The topic may be set inside any template AIML is not exactly the same as a simple database of questions and answers The pattern matching query language is much simpler than something like SQL But a category template may contain the recursive srai tag so that the output depends not only on
105. sa hierarchy consists of a set of assertions about binary predicates like isa dog carnivore isa dog pet isa cat carnivore isa cat pet isa carnivore animal and so on The complete set of these assertions used in this experiment appears below In addition to the assertions there is a general associative rule something like isa X Y isa X Z isa Z Y Through the associative rule the program can reason about isa questions and answer even ones that are not explicitly stated The intuition behind the rule is X is a Y provided that there exists a Z such that X is Z and Z is Y Through the associative rule the bot can reason that since a dog is a carnivore and carnivore is an animal and an animal is alive then therefore a dog is alive Although the preceding Prolog fragment is syntactically correct unfortunately it produces an infinite loop This appears to be a side effect of the order of evaluation in Prolog s built in backtracking The infinite loop is avoided by introducing a new predicate isb X Y defined as follows isb X Y isa X Y isb X Y isa X Z isb Z Y IS Finally the Prolog engine tries to find not just one but the set of all solutions for a particular problem In the interactive console mode the user would normally step through all solutions with carriage return In our non interactive mode we want only to find out whether the set of solution
106. show the compression power of the Graphmaster graph at its best 25 Plotting the patterns from Atomic aiml we are viewing 1 4 of the Graphmaster graph The yellow circles indicate terminal nodes and the magenta are nodes with a template 26 The idea is to plot the root of the graph as a spiral because it has the most branches With our 24 000 categories the number of choices for the first word in this plot is exactly 2001 so the spiral has 2001 1st level branches The plot shows the graph down to level 3 with words for the first two levels 27 The plot shows all of the branches of the graph with 24 000 categories loaded The spiral itself is the root with a branching factor of about 2000 The number of choices for the first word in an input sentence The trees show the average branching factor for the second word is around two although there are a few dense bushes representing patterns starting with I YOU HOW WHAT WHEN WHERE WHO and a few others 28 v gt The plot shows 1 4 of the categories from the file Atomic aiml The color scheme shows each level of the graph as a different color Most categories in the ALICE Brain have lt that gt and lt topic gt both equal to The corresponding redundancy in the Graphmaster subgraphs is illustrated in the lower right 29 This high resolution 1024x768 plot shows the ALICE Brain with 24 637 categorie
107. ss via the natural language interface Human start Bot what is your name Human rich Bot Nice to meet you rich Are you a man or a woman Human male Bot Your gender is recorded as male What is your birth date 80 Human 1960 Bot Being born on 1960 is very lucky Who is your mother Human minty Bot Thank you for the information Who is your father Human stan Bot Great So far we have Name rich Gender male Birth date 1960 Mother minty Father stan Do you have any corrections now Eventually the bot tries to get more of your personal biographical information number of siblings favorite color favorite movie location etc Then we add a personality test The Interviewer tries to classify the botmaster s personality type in order to select a basic bot personality profile in other words an AIML set customized for a specific personality type The Interviewer will also ask the botmaster to enter some pickup lines to customize the AIML personality In case it is not obvious the goal here is to digitize the botmaster s personality The results of this interview can be dumped into a new bot personality And boom in about 2 hours your personalized custom chat bot is complete and ready to chat for you Someone asked what other language acquisition system is there besides learning Well botmaster training is one example have argued that a group of funded botmasters could reach the goal of A I
108. star DATE lt date gt NUMBER none PAYEE unknown CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE unknown CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt PAID WITH CHECK lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER lt star index 2 gt PAYEE UNKNOWN lt srai gt lt template gt lt category gt category gt lt gt 1 PAID DOLLARS WITH CHECK NUMBER lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER lt star index 3 gt PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I WROTE A CHECK FOR TO lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER UNKNOWN PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category pattern l CHARGED DOLLARS ON MY CARD lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED DOLLARS ON MY ATM lt pattern gt lt template gt
109. star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt lt srai gt lt template gt lt category gt category pattern l CHARGED DOLLARS ON MY ATM CARD FOR AT lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category pattern l CHARGED DOLLARS ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON MY FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER non
110. star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt gt SPENT DOLLARS ON ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt 76 category gt lt pattern gt I PAID FOR ON pattern template srai DEBIT ACCT wallet AMT star DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category pattern l BOUGHT FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star index 2 gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category
111. sts then the problem is to match the image with one of the terrorists This case shows that there may be a special category C indicating no match In the case of AIML the input space P consists of all 3 tupes of input that topic strings in normalized form The AIML categories partition the pattern space into disjoint regions determined by the order of the matching function If the only category is the default one with pattern that topic then it both partitions and fills the pattern space Every input matches that pattern Adding one more category pattern HELLO pattern and that topic partitions the input space into two regions those that match this new category and those that match the default Can probability statistics weights neural networks or fuzzy logic improve bots Statistics are in fact heavily used in the ALICE server but not in the way you might think ALICE uses Zipf Analysis to plot the 39 rank frequency of the activated categories and to reveal inputs from the log file that don t already have specific replies so the botmaster can focus on answering questions people actually ask the Quick Targets function Other bot languages notably the one used for Julia make heavy use of fuzzy or weighted rules We see their problem as this the botmaster already has enough to worry about without having to make up magic numbers for every rule Once you get
112. sub graph rooted at the child node linked by X using the tail of the input the suffix of the input with X removed If no match was found try Does the Nodemapper contain the key If so search the sub graph rooted at the child node linked by Try all remaining suffixes of the input following X to see if one matches If no match was found go back up the graph to the parent of this node and put X back on the head of the input For completeness there should also be a terminal case If the input is null no more words and the Nodemapper contains the template key then a match was found Halt the search and return the matching node If the root Nodemapper contains a key and it points to a leaf node then the algorithm is guaranteed to find a match Note that At every node the has first priority an atomic word match second priority and a match lowest priority The patterns need not be ordered alphabetically only partially ordered so that comes before any word and after any word The matching is word by word not category by category The algorithm combines the input pattern the that pattern and the topic pattern into a single path or 21 sentence such as PATTERN that THAT topic TOPIC and treats the tokens that and topic like ordinary words The PATTERN THAT and TOPIC patterns may contain multiple wildcards The matching algorithm is a highly restricted version of de
113. t MUCH MONEY WALLET pattern template srai BALANCE ACCT WALLET srai template category gt category gt lt patterns gt HOW MUCH CASH _ HAVE pattern template srai x BALANCE ACCT WALLET srai template category gt category gt lt pattern gt WHO IS MY LANDLORD pattern template get name landlord gt lt template gt lt category gt category gt lt pattern gt WHAT IS MY RENT lt pattern gt lt template gt lt get name rent gt per month lt template gt lt category gt lt category gt lt pattern gt MY RENT IS lt pattern gt lt template gt lt set name rent gt lt star gt lt set gt is the market rate lt template gt lt category gt category pattern MY RENT IS DOLLARS PER MONTH lt pattern gt lt template gt lt srai gt MY RENT IS lt star gt lt srai gt lt template gt lt category gt category pattern MY RENT IS PER MONTH lt pattern gt lt template gt lt srai gt MY RENT IS star srai template category gt category pattern MY LANDLORD IS pattern template lt think gt lt set name landlord gt lt person gt lt set gt lt think gt Rent is theft lt template gt lt category gt category gt lt pattern gt MY CHECKING ACCOUNT BALANCE IS lt pattern gt lt template gt lt srai gt INITIALIZE BANK ACCT checking AMT star srai template category gt category pattern MY CIT
114. t star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt PUT DOLLARS ON CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT MY CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt gt PAID DOLLARS FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT wallet AMT lt star gt DATE lt date gt NUMBER none PAYEE
115. t star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT ON MY CARD ON lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I CHARGED MY CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 4 gt CATEGORY l
116. t star index 5 gt CATEGORY lt star index 6 gt NOTE NONE lt srai gt lt template gt lt category gt category gt lt pattern gt DEBIT ACCT AMT DATE NUMBER PAYEE lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star gt AMT lt star index 2 gt DATE lt star index 3 gt NUMBER lt star index 4 gt PAYEE lt star index 5 gt CATEGORY UNKNOWN lt srai gt lt template gt lt category gt category gt lt pattern gt DEBIT ACCT AMT DATE NUMBER lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star gt AMT lt star index 2 gt DATE lt star index 3 gt NUMBER lt star index 4 gt PAYEE UNKNOWN lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED ON CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT star index 2 gt AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 4 gt lt srai gt lt template gt lt category gt category gt lt pattern gt DOLLARS ON CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT lt star index 2
117. t to create a chat robot Not difficult If you can write HTML you can write AIML Artificial Intelligence Markup Language Here is an example of a simple but complete chat robot in AIML category lt pattern gt lt pattern gt lt template gt Hello lt template gt lt category gt lt aiml gt The tags lt aiml gt lt aiml gt indicate that this markup contains a chat robot The lt category gt tag indicates an AIML category the basic unit of chat robot knowledge The category has a lt pattern gt and a lt template gt The pattern in this case is the wild card symbol that matches any input The template is just the text Hello As you may have guessed this simple chat robot just responds by saying Hello to any input You can get started with AIML knowing just the three tags category pattern and template much like you may have 37 started with HTML knowing only a img and lt h1 gt Does ALICE learn The model of learning in ALICE is called supervised training because a teacher the botmaster always plays a crucial role The alternative unsupervised training is complicated in an open environment like the Web The problem is that clients are untrustworthy teachers and forever try to fool the robot with untrue assertions Does ALICE think It depends on what you mean by thinking The most fascinating responses from ALICE arise when she says som
118. tar index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT DOLLARS ON MY ATM AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY ATM CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I DOLLARS ON MY FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY ATM AT FOR 65 lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index
119. te NUMBER none PAYEE unknown CATEGORY star srai template category gt category pattern MY RENT IS A MONTH lt pattern gt lt template gt lt srai gt MY RENT IS star srai template category gt category pattern MY RENT IS DOLLARS MONTH pattern template srai MY RENT IS lt star gt lt srai gt lt template gt lt category gt category pattern MY RENT IS DOLLARS pattern template srai MY RENT IS lt star gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY ATM FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 2 gt CATEGORY lt star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt CHARGED DOLLARS ON MY CARD ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt gt 1 CHARGED ON MY ON FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category gt lt pattern gt l CHARGED ON
120. tegory pattern DEPOSITED TO pattern template srai CREDIT ACCT star index 2 gt AMT star lt srai gt lt template gt lt category gt category gt lt pattern gt _ LOANED ME DOLLARS pattern template srai CREDIT ACCT wallet AMT star index 2 gt DATE date NUMBER none PAYER lt gt CATEGORY loan NOTE repay lt person gt srai template category gt lt category gt lt pattern gt _ LOANED ME pattern template srai CREDIT ACCT wallet AMT star index 2 gt DATE date NUMBER none PAYER person CATEGORY loan NOTE repay person lt srai gt lt template gt lt category gt category gt lt gt 1 CHARGED DOLLARS ON MY CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category pattern l CHARGED ON AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY star index 3 gt lt srai gt lt template gt lt category gt category pattern l CHARGED ON MY CARD AT FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 2 gt CATEGORY s
121. there try the patterns beginning with The only final complication is that the and appear in patterns that begin with 0 9A Z If your input matches some pattern prefix then try the patterns with that prefix followed by first then the ones with that prefix followed by some word finally try the ones with the prefix followed by THE ALICE BRAIN PICTURE GALLERY we gt c rr 5 22 di 7 t TU t EJ 2 h own 23 The eye shaped log spiral plots all 24 000 categories in the ALICE Brain The spiral itself represents the root The trees emerging from the root are the patterns recognized by ALICE The branching factor for the root is about 2000 but the average branching factor of the second pattern word is only about two Those acquainted with my earlier work on logmap sensors and cortical algorithms for visual processing may notice the similarity with the Graphmaster plot This similarity is in my opinion no coincidence The same cortical architecture that enables real time attention based visual processing in fact be applied to linguistic processing as well 24 CBA are sph grass y T This plot shows 1 4 of the patterns in Srai aiml or about 2000 patterns total The four big trees in the lower right are all the patterns rooted at the words WHAT WHEN WHERE and WHO respectively from right to left These areas
122. though the pattern contains a wildcard 41 Another default category of little use for targeting is the one with pattern CALL ME All of the name telling categories such NAME IS MY REAL NAME IS and the one with that WHAT CAN CALL YOU lt that gt reduce to the CALL ME category Unless you are interested in writing special responses for many different people s names then the CALL ME category is not a good source of targets Finally if none of the previous cases apply the program considers whether the matched pattern contains a wildcard If the pattern is atomic the category is not likely to be a good source of targets It could be if we considered that as in the first case or topic but that would be too advanced for this algorithm If the pattern contains a wildcard then the category is likely a good source of targets If the matched pattern contains a wildcard the suggested new pattern is generated as follows Align the input sentence with the matched pattern Create a new pattern by using the words from the frst pattern plus one more word from the input SOME EXAMPLES 1 Suppose the topic is READY and the following conversation fragment takes place Robot Oh Client WHAT IS PIZZA Robot have to process that one for a while The input matched the category with pattern WHAT IS A lt pattern gt generating a default response The value of that is OH Browsing the tar
123. tly the same format except that the DEBIT transaction has a payee where the CREDIT has a payer In either case the program stores the transactions in a comma delimited format making for easy import into other financial applications Specifically the transaction record format is a comma delimited string of the form DEBIT account amount date number payee category note id CREDIT account amount date number payer category note id The account field identifies the account as checking savings wallet or the name of another account The amount is indicated in dollars as a nonnegative integer The date includes a time stamp to the resolution of seconds The number field usually indicates a check number or transaction number if any Payer payee indicates the source destination accounts for the transaction The category field helps create a budget report later showing how our income and expenses break down by categories Such as rent medicine food utility bills and so on The security of the system is limited to recording the client id in this case localhost localdomain If the bot is online care should be taken so that unauthorized clients cannot initiate personal financial transactions The records do not include a unique identifier or row number Uniqueness of the records is only guaranteed by the timestamp using the AIML lt date gt function This is fine if the transaction rate is always less than on
124. transaction recorded in the gossip file Ideally the AIML script should be tightly integrated with a real spreadsheet application For now we use a simple shell script to stand in for the real application The script uses awk to calculate sums Its use of temporary files is unfortunate the account balance shell script This example is for instructional purposes only Usage balance lt account gt rm f temp echo 0 gt credit temp echo 0 gt debit temp echo Your balance is grep DEBIT home alicebot data gossip data grep awk F print 2 n 2 print tail 1 gt debit temp grep CREDIT home alicebot data gossip data grep awk F 2 2 print tail 1 gt credit temp cat credit temp debit temp awk print 1 n 1 print tail 1 AN AIML SPREADSHEET We have demonstrated the feasibility of a simple personal finance application with AIML Despite its limitations the program is adequate for a small set of personal finance needs To make this application more robust we should Create a realistic double entry bookkeeping system b Add a unique identifier to each transaction 63 C Integrate the program more closely with a real financial application d Develop an expanded set of AIML patterns to capture a wider variety of transaction statements including dollars and cents international currencies interest payments and taxes e Enha
125. ttern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I PUT ON ATM CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM CARD FOR lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category gt lt pattern gt I SPENT DOLLARS ON MY ATM FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT lt star gt DATE lt date gt NUMBER none PAYEE lt star index 3 gt CATEGORY lt star index 2 gt lt srai gt lt template gt lt category gt category pattern SPENT ON MY ATM CARD FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checking AMT star DATE lt date gt NUMBER none PAYEE star index 3 gt CATEGORY star index 2 gt lt srai gt lt template gt lt category gt category pattern SPENT ON MY ATM FOR ON lt pattern gt lt template gt lt srai gt DEBIT ACCT checki
126. tween a client and the chat robot with a Prolog based commonsense reasoning system Client Is Rich alive Robot Yes Client Is a boy a mammal Robot Yes Client Is a dog a vegetable Robot No 48 Client Is a cat a pet Robot Yes Client Is a penguin food Robot No Client Is a human a mammal Robot Yes Of course such a model of reasoning is by itself severely limited One obvious flaw is that the bot will always answer No for any out of domain questions Client Is a frog an amphibian Robot No Client Is a watch a clock Robot No Client Is a horse an animal Robot No This problem could be approached by adding code to the Prolog program to make sure that the elements of the query are also elements of the domain of isa Another problem is that the pattern IS A A may match other inputs for which the PSYCH program is totally inappropriate Is a dollar a piece a good price 5 gram an ounce or a pound bigger Is a good thing ACTIVATING PROLOG FROM AIML The experiment described here demonstrates how to activate a simple Prolog program using the system tag of AIML It shows that quite a bit of commonsense reasoning can be added to an AIML chat robot robot without proposing or adding any new AIML features The PSYCH program demonstrated here is currently limited to only isa relations but more commonsense predicates can be added easily For example some simple Prolog
127. up 10 000 categories like ALICE you don t want to think about more parameters than necessary Bot languages with fuzzy matching rules tend to have scaling problems Finally the bot replies are not as deterministic as you might think even without weights Some answers rely on random to select one of several possible replies Other replies generated by unforeseen user input also create spontaneous outputs that the botmaster doesn t anticipate Can have a private conversation with ALICE The ALICE server logs and records all conversations Even the ALICE Applet tries to transmit conversation logs back to the originating server You can have private conversation with ALICE however if you download Program B D or one of the other free ALICE downloads to your own computer and run it there Running on your machine the server stores all the conversations locally How do l catch all occurrences of a keyword in the input You only need pattern KEYWORD pattern patterns KEYWORD lt pattern gt pattern KEYWORD lt gt patterns KEYWORD lt gt Use lt srai gt KEYWORD lt srai gt for all but the first template Place the desired response in the first template What is the exact use for and When to use one and when another They are exactly the same except that _ has priority over any word and any word has priority over The meaning is match one or more words
128. w access to each customer Pandorabots will be adding more services in the near future At the time of this writing there is no charge for these additional services We expect future services to be compatible with this service yet we can make no guarantees Also as this hosting site is free of any charges please be aware that we make no guarantees or warranties of any type How can Sign Up Customers Recruiting and Persuading Customers Use your imagination to recruit new customers To get started consider these ideas 1 e mail to promote your site 2 advertise your robot from a free demonstration robot supplying free sample interactions 3 advertise on other botmaster s robot pages 4 paying other botmasters to advertise your robots 5 or use other botmasters robots to advertise your pages Advertisements can be placed on pages of particularly popular robots Embed advertising messages into your robots and sell the service to other companies and botmasters Use your imagination When a customer makes a purchase with PayPal either by clicking a button on a web page or following a link in an e mail the PayPal system will alert you the seller by sending you an e mail You can also go to the PayPal site and log into your account to check your daily activity Any new orders appear in the list of daily transactions in your account Once you link your checking account to PayPal you will have the option to withdraw funds and trans
129. y 35 category gt pattern NO pattern that l UNDERSTAND lt that gt template srai YOU DO UNDERSTAND lt srai gt template lt category gt illustrates the use of that with lt srai gt The client input matches simply No What did the robot say that made the client say no If it was understand then this category formulates a response with lt srai gt YOU DO NOT UNDERSTAND lt srai gt which in turn activates another category with the markup pattern YOU DO NOT UNDERSTAND lt pattern gt This category responds do so understand It all makes sense to my artificial mind The AIML srai tag simplifies and combines four important chat robot operations Maps multiple patterns to the same response Reduces a complex sentence structure to a simpler form Diminishes the need for multiple wildcard input patterns Translates state dependent inputs into simpler stimulii In some sense srai is a very low level operation but its simplicity captures a wide range of typical chat robot functions CHAPTER IV Botmaster Questions and Answers What is the goal for AIML AIML Artificial Intelligence Markup Language is an XML specification for programming chat robots like ALICE using program B The emphasis in the language design is minimalism The simplicity of AIML makes it easy for non programmers especially those who already know HTML to get started writing chat robots
130. y meaningful the number of things people actually do say is surprisingly small Steven Pinker in his book How the Mind Works wrote Say you have ten choices for the first word to begin a sentence ten choices for the second word yielding 100 two word beginnings ten choices for the third word yielding a thousand three word beginnings and so on Ten is in fact the approximate geometric mean of the number of word choices available at each point in assembling a grammatical and sensible sentence A little arithmetic shows that the number of sentences of 20 words or less not an unusual length is about 10 na Fortunately for chat robot programmers Pinker s calculations are way off Our experiments with A L I C E indicate that the number of choices for the first word is more than ten but it is only about two thousand Specifically about 2000 words covers 9596 of all the first words input to A L 1 E The number of choices for the second word is only about two be sure there are some first words and You for example that have many possible second words but the overall average is just under two words The average branching factor decreases with each successive word 19 We have plotted some beautiful images of the A L I C E brain contents represented by this graph http alice sunlitsurf com documentation gallery More than just elegant pictures of the A L I C E brain these spiral images see more

The Elements of AIML Style

Contents

Download Pdf Manuals

Related Search

Related Contents