OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22;...

24
OB {3l WED 14 JUL 76 Page 2 <AVRON>HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring of the field inspired by our discussion with Cordell Green, Int roduct lon phi losophy relation to society history sources (Further) Intelligence (active Ingredients) Representation of Knowledge Problem Solving search - as a way of representing problem solving armingI anning Reasoning by Analogy Pattern Matching Learning Information Processing Psychology Areas where we try to create Intelligent Behavior Games Natural Language understanding Robots vi sion Speech Understanding Automatic Programming Theorem Proving app! Icat lons, of A] In the real $1.0 i-Kp i^>c-a\^o worl d AI tools and techniques (How we do our magic) Languages & systems Search - as a method Generate and Test Exhaustive Search Heuristic Search Knowledge-based heuristic search Specific Representation Schemes theorem proving (resolution) automatic programming (someday)

Transcript of OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22;...

Page 1: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

OB {3l WED 14 JUL 76 Page 2 <AVRON>HAI,HANDOUT-MAY22; T

gfc

"

"

Notes on the Structure of AI

A restructuring of the field inspired by our discussionwith Cordell Green,

Int roduct lon

phi losophyrelation to societyhistorysources (Further)

Intelligence (active Ingredients)

Representation of KnowledgeProblem Solving

search - as a way of representing problem solvingarmingI anning

Reasoning by AnalogyPattern MatchingLearningInformation Processing Psychology

Areas where we try to create Intelligent BehaviorGamesNatural Language understandingRobotsvi sionSpeech UnderstandingAutomatic ProgrammingTheorem Provingapp! Icat lons, of A] In the real$1.0 i-Kp i^>c-a\^o worl d

AI tools and techniques (How we do our magic)

Languages & systemsSearch - as a method

Generate and TestExhaustive SearchHeuristic SearchKnowledge-based heuristic search

Specific Representation Schemestheorem proving (resolution)automatic programming (someday)

Page 2: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

08131 WED 14 JUL 76 Page 1 <AVRON>HAI,HANDOUI-MAY?2; I

"

"

"

May 22, 1976

the HA! Semantic Net

Basic Idea

A relatively naive user should be able to probe the Handbookto become familiar with any aspect of AI, at least to the point where hecan be sent to the original literature with some hope of understanding it,

Types of Nodesoverviews

3+ pagescover a sub-fields ideas, different approaches, failuresreferences and articles on specific topics

Articles

2-3 pagescover a specific topic (technique or program)references

meta-nodes

2«3 paragraphsexplain the relationship between several articles

Glossary entries

one-Hners defining terms and pointing back to thearticles that ust them f

SißLioCpdAc-t^ 6+rrt.fcTypes of relations

is an example of (SHRDLU, nl programs)is a sub-category ofrefers to explicitlyrefers to Implicitlyglossary terms

(semantic net* representation)(production systems* productions)(LISP, list Processing Languages)

references (annotated)prerequisite readingbrought to you by

(wfiOQMC, ai/7"Jwrought to you by -CAi ,—(MoCarthyy—Arpa>) (sy&o^v >o/A

IS uiVo(fri,A root}) /» (s6M2Jt,vf , c/AteS FAUXfIUhMk/ ad. £(f/I*Vv<LTIOA>S j WC/tO)

| P<£4

Page 3: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

" " "

Page 4: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 5: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

-r

"

«-£ ££#&£=*

mg*. tot2- 7Z& S&Afte-H 7705. t saZMul>l-1

Jp/hc&

(Z£c*t44 £apuctt4 fi^io/o/^ AfeAftrs 3S7-fcn» spfhL^ Ptt>'n7i -/Zv-P&r~'

PWZtfZStscAsi MuSSoa)

Co AJ7JA.T S^-fr^Z, SpA-z-l^" /- ,***. ?**tr^ r^"1

*r-pfto&lJf^ &H>vCJrzO*J Mill*&*

P^liA^VO P&3&L&

*5/r-gAftiV\p^L OR

/S/l 0-X T=>OL

fao/oa fft-ftf'T^

Sxtrpa. <** ftq***t*r**"> *******

${UH«<XS *frZL T^©l_

Sp/H^. <<TSv4^r/

<T*pi,]A^>o /OoD£.

Page 6: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

"

"

Page 7: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 8: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

finished printing <KIDBIP>TSK-IN-TNQ f ;37 PAGES, 3675' CHARACTERS PRINTED AT 8000,1 CHARACTERS / SECOND

Page 9: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

tggf^

"

Page 10: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 11: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 12: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

"

"

fAfi-ir

UptA-yo^

9.IIHAS to (/Zxpuar-UAJ - o£*4eeD

/) AJfirvS. VSZB- BZ. fiBLZ To 7=¥2o/S'£. 77*1, tyS7X^ (i» &>fu**J

7>/30p572-X7 SA

. 0/z. TUpWl-

uori-o z>D*3t* aa .

<_-*, /S A9*J S.xSWpl-Z' £>'z £-

tL—5> &£ («#2#/o<^£-4 (^oP-OifitO 8u J

]-*****■ n » fimßar

l^rcJ- *««> * **BB '„*> ,t fi f°°a 'a**r "

Page 13: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 14: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

'PcAh^mtpC. S<2/W2-cJ-'

"fOOL^

" T&&& fZlpPtUT*****

(gaff£C\t+)CA. - rv)^7/ )

/hbS

"

Page 15: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 16: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

(Bl MTSrf.&'hzZ (#/ A>^<-££>£"£ £jc£>z/2JT

cy-zxs

*IPZ-MAo. vA-u. ftoe>f/ZJ**<s

tffiP* **TC-»»'fi

fyizc^ nt* figz f^TvfLl^

\SN9ujf±ry^ Z?////o£" SU,&Lr-&Lil4,

A/1 TSUtft-fiC C^TtetUl*,

P<-A^>]i-/^lT(U-r^ r> i/ocAgvoA^tU ciWS

_z - rA\)c<L SWTvi

"

Page 17: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring
Page 18: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

" tx /? SZ/HLctf

A TrJ

/v\Y<l/a>

P/2oPi/C r?oA> Sy i 7"**k i

ZpiijLJt -vnt>?M> r***oi>/^ -Oo LBytUy>

yurHT&i^

"

"

Page 19: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

H133 THU 13 MAY 76 Paae 1 <AVRON>POINTE,BEST-I-«C0UID;1

[frequency search on DSKjPQINTE1 a1 acoustic-Processing

r^px i ad-hoc heuristics1 advice taker1 alpha beta search techniqueI alpha-beta procedure2 alpha-beta tree searchI analogyi ancestory strategies in resolution theorem proving1 ancestry filtered strategies1 end-or graphs1 and/or trees1 assertions1 associative data bases1 atn grammari augmented transition networksI automatic program synthesis3 automatic programming3 automatic theorem proving1 backtracking1 blndlng/access/control llnke1 bledsoe's systems\ b) lnd searchI blind-search methods3 breadth f i rst searchI breath-first1 bridge

"1 case frames1 checkers2 Chesschessl chess programs1 clowes1 cluster analysis1 concept formation1 concept-formation3 connlver1 connlver1 context-free grammar1 contour following1 control structurei control structures1 d, a, waterman's machine learning of heuristics2 d,a, waterman's machine learning of heuristicsq demons1 dendralI depth firstq depth first search3 depth first search techniquet depth search techniques1 dragon1 edge enhancement1 epam2 evaluation function

"2 evaluation functionsX evensevensI feigenbaum3 formal grammarsI frame problem1 frame-systems

Page 20: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

14133 THU 13 MAY 76 Page 1,1 <AVRON>POINTE,BEST-I-COULD; 1

lb

32

2

?

" 533

?

"

game playinggame-playinggelernter's geometry provergogpsgrammarsgrammars of various typesgrammmarguzmanhackerhearsay 1hearsay iiheterarchleheuristic searchheuristicshierarchical data baseMil climbingHuf fmanIbm systemInductive InferenceInductive-InferenceInferent l al memoryInformal program specification methodsInformation retrievalInter! lspInterl lsp1p1»vknowledge from multiple sourcesknowledge representationknowledge representation using the predicate calculusknow ledge-represent at lonlanguages overviewlearning by building structural modelslearning by debugginglearning by parameter adaptationlearning by paramteter adaptation1 ne descriptionsi isplist processing languagesItlunar systemmachine representation of knowledgemachine visionmechanical theorem provingmedical computingmedical decision makingmeteormjn-max tree searchminimaxminimax procedureminskymodelmodular representationmove generatorsmycinnatural languagenatural language articlesnatural language parsingnatural language understandingnevlns' work

Page 21: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

14833 THU 13 MAY 76 Page 1.2 <AVRON>pOINTE,BEST-I-COULDf 1

?31

" 1A11111«11311111111I2t1

" 12?411121131l321111611

H

I111

noise removalordered searchother speech understanding systemsparallel processesparameter adaptation of linear functionsparameter adaptation of nonlinear functionsparameter adjustment of linear functionsparambdul at ionparserspattern directed computationpattern directed invocationpattern directed procedure invocationpattern matchingpattern-matchingpicture codingplannerPi annerPlanner and connlverplanningpokerpredicate calculuspredicate-calculus representationproblem behavior graphsproblem domainsproblem reductionproblem reduction representionproblem solvingproblem solving by theorem provingproblem solving via theorem provingprocedural representationprocedural representation of knowledgeproduction rulesproduction systemsproductionsprogram synthesisprogram synthesis from examplesprogram synthesis from tracesprogram synthesis via theorem provingProject macprotocol analysiMSqa4ql Ispql lap programquestion answeringquestion answering via theorem provingquestion-answeringqui 1 1 ianregion-growingrepresentation of knowledgeresolution theorem provingrobot problem solversrobot problem solvingrote learningSamuel's checker playing programsscene-analysissdc systemsearch and/or graphssearch applied to state spacessearch strategies

Page 22: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

1«|33 THU 13 MAY 76 Page 1,3 <AVRON>POINTE, BEST-I-COULD;!

sIH

"utlon theorem

proving

"

4 search techniquessearch trees

semantic netssemantic strategiesshape descriptionshrdlushrdlusignature and move phase tablessignature and move-phase phase tablessltuatibnal-f luentsslipsolution theorem provingspatial filteringspeech understanding systemsspeech understanding-acoustic processingspeech-understandingspeech-understanding-overvlewspeech! issri systemstate space representationstate space searchstate-space paradigmstrip'ssussmansyntac- ordering strategies 1n resolutionsyntactic and semantic refinement strategies In resol

5

syntaxsynthesis from examplestheorem prover implementationtheorem provingtheory formation In program synthesistransformational grammartransformational grammarstree searchwaltzwal t*wood's lunar systemword-associationyorlek wilks' workalgorithm

1 lap

Ithere arp 319 words on the file DSKiPOINTE,1221 distinct words,(

Page 23: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring

10:37 THU 15 APR 76 Page 15 <AVRON>AIH.OUTLINE; 1

XII

"XIII

"

"

where do these go ?

Reasoning by analoTy

J-** el 1 1 gence augmentation

Chess

Learning a^d Inductive Inference

OverviewSamuel Checker program

" [niton

Pattern extrapolation probl ems--Slmon,Overview oi Induction

done 14)

done 151

done C5)

NEED [91

NEED [53

done t2)

WEED [53

Page 24: OB WED JUL 2 HAI,HANDOUT-MAY22; · 2015. 10. 24. · OB{3l WED 14 JUL 76 Page 2 HAI,HANDOUT-MAY22; T gfc " " Notes on the Structure of AI A restructuring