"

"

"

Page 1 AIH, OUTLINE;!

AI HANDBOOK OUTLINE

I, INTRODUCTION

A, Intended Audience

This handbook is intended for two kinds ofaudience; computer science students interested1n learning more about artificial intelligence, andengineers in search cf techniques and ideas thatmight prove useful in applications programs.

B, Suggested Style For Articles

The following is a brief checklist that may providesome guidance in writing articles for thehandbook. It is, of course, only a suggested list,

i) Start with 1-2 paragraphs on the central idea orconcept of the article. Answer the question "whatis the key Idea?"

il) Give a brief history of the invention of theidsa, and it s use in A.l,

iii) Give a more detailed technical descriptionof the Idea, Its implementations in the past,and the results of any experiments with It, Tryto answer the question "How to do It?,

iv) Make tentative conclusions about the utility andl imi tat lons of the idea if appropriate,

v) Give a list of suitable references,vl) Give a small set of pointers to related concepts

(general /overview articles, specific applications, etc.)

C, Coding Used In This OutlineThis outline contains a list of the major_areas of artificial Intelligence covered in the

Ibook, At, the lowest level, the outline showsarticle titles either contained or needed. Inthe case of an article that is needed, the notationNEED*, follows the proposed focus of thethe article, where Ms a number in the interval[0(101. Low numbers indicate little expecteddifficulty with the article, whereas high numbersindicate a Potentially difficult article. For

10:34 THU 15 APR 76

* 10:35

"*****

"

"

Page ? AIH, OUTLINE;!

iffipjtf an article on a specific system, where onlyI minimal emount of reading is required wouldrate approximately 4, whereas an overview articlewould likely rate 8 or greater, In the case of"ftieltl which alrsady exist in the handbook, thenotation donet*3 is used, where low numbers indicatethat the article needs only minor mooi f icat ions, andhigh numbers indicate that major modifications arerequired, For example, repair of typographicalerrors and wording could be expected to rate 0-2,Correction of errors in the article might rate 3-6,and major rewrites which require considerable readingwoul d 1 1 kely rate 7-10,

It should be noted that the real difficulty involvedin writing an article is highly dependent on the a prioriknowledge of Its author.

This has been a preamble. Now for some areas not covered

10

II

"

"

"

;35 THU 15 APR 76 3 I . I 5 AIH, OUTLINE;

HEURISTIC SEARCH

A, Heuristic Search Overview NEED [9]Mgor.tMfc presentation of "heuristic search" procedure.Heur.st.cs for choosing promising nodes to expand next,

heuristics for choosing operators to use to expand a node.using heuristics to choose relevant heuristics.

Pervasive character of the combinatorial explosion,Arguments (both formal and Intuitive) supporting the use of

heuristic search to muffle this explosion.Formal. Completeness of A*; Knuth's recent work on

aloha-beta search.Opportunities for future research

"there do heuristics come from?(see Simon's current work; meta-rules; met a-meta-, , .?)

ribd.lytf.fl heuristics based on experiencesCsee Berliner's current work)

Working with symbolic? rather than numerical, values for nodesCoding heuristics as production rules

(e.g.: view Mycin as a heuristic search)Situations NOT suited to attack by heuristic search

Typically; non-exponential growth process; no search anyway(e.g.* finding roots of a quadratic equation)

Identity oroblemsDj 33 ulsfng Heuristic Search as something elseDisguising something else to appear to be a Heuristic Search

P. Search Spaces

1. Overview NEED .83The concept of a search space; how a search spacecan be used to solve .some) problems; differentrepresentations, different spaces

2, State-space representation done [_>.[2 articles exist h ere, which ought to be unified!

.». Drobl #m-reduct ion representation done __>_

4, ANO-QR trees and graphs done [41

li***^> A*\

105 35 THU 15 APR 76 Page '4 AIH,OUTLINE; 1

":

"D

"

"

"Bl ind" 'Search Strategies

1, Overview NEED [51

done .23done (23

NEED £6J

2, Breadth-first searching

3, OePth-first searching

'4. Bi-directional searchingdiscuss heuristics. Ml articles by ira Pohl

5, Minimaxlng done .3.

done .316, Alpha-Beta searching

Using Heuristics to Improve the Search1, Overview done 171

The idea of a heuristicThe idea of a heuristic evaluation functionsavings in change of representation.

2, Best-first searching done (43 .CQr .ered-search) but need to add; Martelll'swork (ask Nils for a dra*t of this)speech rec: IJCAI-3 (Paxton), Reddy's book

3. Hill climbing done [33

4. Means-ends analysis done (33

5, Hierarchical search, planningAbstrips (Sacerdoti)

in abstract spaces NEED (43

6, Branch and bound searching done (43

NEED (431 . searchingHarris - AI Journal

AIH, OUTLINE;i 0 5 35 TH J 15 A^R 76 Page 5 1

E

"

"

"

Programs employing (based on) heuristic search

1, Overview NEED (73Comparison of systems. Results & limitations,(This first article should be written later as anintroduction to the following articles,)

2, Historically important problem solvers

a) GPS

b) Strips

c) Gelernter's Geom, Program

NEED [43

NEED (43

NEED (33

10

______ ITI

"

11

"

35 THU 15 APR 76 Ii 6 AIH, OUTLINE

Natural L^Tnuag*

A. Overview

1, early machine translationFailures of straight forward approaches

done [53

2, History and Development of N,L, NEED (83Main ideas (parsing, representation)comparison of different techniques, mention ELIZA, PARRY,Include Baseball, Sad Sam, SIR and Student articles here,see Hlnograd' s Five Lectures, Simmon's CACM articles.

_). Representation of Meaning(see section VII -- HIP)

C. Syntax and Parsing Techniques

1, overview of formal grammars2, augmented transition nets, WoodsS. ShricMu'l parser (systemic grammars)4, Case Grammars Bruce (AI Journal, 1/76)5, CHARTS - well formed substrings6, GSP syntax 8 Parser7, H, Simon - problem understandingla transformational grammars

done [33done (33done (53NEED (53NEED [63NEED [63NEED (73done [53

D, Famous Natural Language systems

1. SHRDLU, Kinogred2. SCHOLAR

NEED (53NEED (53NEED (533, SOPHIE

E, Current translation techniques NEED [8]bilks' work, commercial systems (Vauquois)

F. Text Generating systems NEED [8]Goldman, Sheldon Klein, Simmons and Sloan (In S&C)

10:35 THU 15 APR 76 Page 7 AIH, OUTLINE; 1

J^ IV, A I Languaget

Early list-processing languages done (33

f overvi ew articlelanguage? 1 1 _)" SLIP, LISP, IPL 3r*eß © Q^»P to^TT «ideas: recursion, list structure, »ltT#__*. fvr«s ~ j0, Ou*rvlew of current H st-processlng c^mf* t,s '14J?g_ieset NEED (731, Control structures, what languages they

are in and examples of their use.NEED (63

BactracMng (parallel processing)Demons (pseudc- 1 nter rupt s)Pattern directed computation

2, Data Structures (lists, associations,bags, tuples, property lists,,,,)

NEED £53

" Once again, examples of their use, Is important here, » \3. rfi-Hnte tAPt-Ukt^g (*»**i»>»*,. 1, LISP, the oasic idee

2. INTFRLISP"v- f ,+l -*3, SUSP (mention QA4)HfiT* 4. SAIL/LEAP?. PLANNER

done (23NEED (53done [33done [23done (23done (23NEED (43NEED [43NEED [43

"

"

"

Page 8 AIH, OUTLINE;!

V, AUTOMATIC PROGRAMMING

A, Cvervi aw done (73

B, Program Specification Techniquesi.e. how does the user describe the program tobe synth?sized?--an overview article Including various methods NEED [93(end done articles on Traces and Examples) andgraphic..! specification. s(&^ .cys**-./.*')

s ~ NEED [93* given a description of the Program in some form,generate the actual program--one article Including classical problem-solvingtechniques, theorem proving, debugging (Sussman'sHacker), Simon's Heuristic Compiler, Prow (Waldeman),in I C, Green' s work.

D, Program optimization techniquesi to turn a rough draft into an efficient

WEED (73

program. See Darlington & Burst-all, Low,

E, Programmer's aids(Interl l»p's DWXM, etc)

NEED [73

NEED [73F, Program verification(TJCAI 3)

10 :36 THU 15 APR 76

10;36 THU 15 APR 76 B |gg ■ AIH. OUTLINE;!

VI. T

■

" .«.B!

n

"

"

.EOREM PPOVING"

_.

m■_ " ■_

_i

m m _» _« -» *

Overvi ew MEED [9]

Resolution Theorem Proving

1, Basic resolution method done (43

2, Syntactic ordering strategies done [23

3, Semantic % syntactic refinement(4, other strategies?)

done [23

Non-resolution theorem proving done £33

Uses of theorem proving

1. Use in question answering done (33

2, Use In problem solving done [63

3,(aA3— PCoU^uS") 179. iftr^o/^ NEED (534, wan-machine theorem proving NEED 163

Predicate Calculus done [53

Page 10 AIH, OUTLINE;!

VII. Human Information Processing -- Psychology

" (see Perry'3outline for details and references)A, Perception MEED [9]An overview of relevant work in psychologyon attention, visuol and auditory perception,Pattern recognition, Applied perception ("ERCFIVER) ,Difficulties resulting from inability to introspect.

■?. Memory and Learning

It 'ia3lc .structures and processes in IPP NEED [93

Short- and Long-term memory, Rehearsal, Chunking,Recognition, Retrieval, recall, Inference and

-»t ion-answerif.fi Semantic vs, episodic memory,?rference and forgetting. Type vs, token nodes

Simon - Sciences of the Artificial?. Overview of memory models, Representation NEED (103

fio ih- j(!fiP'i

'

° 'How to fltt to the airport; A comparison of9 P^cJh^c^^^%J^l^ models.. »tJL *>/ o«*«» a ; Associative memory -model 3#m£ 1. semantic nets NEED [03fcxr>(Zl> Qui Mian CTLC), Nash-Weber (BBN)

Shapiro, Hendricks (3RD, Wood's articleYtf*u>lpJV 1r! 3obrow % Collins, Simmons (S&C)Wtopv&Krf* M€* z, HAM (Anderson ft Bower) NEED [73iUa^tmx 3 » LNRi Active Semantic Networks NEED [63liu*nn? /J< componential analysis NEED [93flsjj T akendoff, Schank (conceptual

dependency), C'ARGtE), G, Miller5. EPAM NEED (5)6. Query languages NEED [7]

wood«s (1968), Ted Codd (IBM SJ)

V. Other representations{^qi $ 1. Production systems done (13p- v 2, Frame systems (Minsky, Winograd) done (7)f° tt, 3. Augmented Transition Networks done [33

,«#

lV 4. Scripts (Schank, Abelson) NEED (73

C, Psychol inguist ics NEED [93

done (73done [33NEED (73

A oros« glossary including;Competence vs. performance models, Phonology vs.

10.36 THU 15 APR 76

10136

"D.

E.

"

"

Page 10,1 AIH, OUTLINE;!

syntax ys. semantics vs, pragmatics, Surface vs,deep str cture, Taxonomic grammars, generative grammars,t rensf ormat i ' : 1 grammars, phrase-st ructure rules,transformation ru]ssr Constituents, lexical entriesParsing vs, generation, Context-free vs. Context-sensitivegrammars, Case systems (e.g., Bruce AI article)

Human Problem Solving -- Overview MEED (8)1, PBG's done [13

done [23NEED [63

2. Concept formation (Moston)3. Human chess problem solving

Behavioral Model ing

1. Belief Systems NEED [83Abelson, McDermott

2, Conversational Postulates (Grlce, TW)3. Parry

NEED (53NEED [53

THU 15 APR 76

. 10:36

VIII,

" A

x

"

c.

"

THU 15 APR 76 Pegs 11 AIH, OUTLINE;!

VISION

Overview MEED [93

This article should discuss the early work in vision;its roots in pattern recognition, character recognition,Pandemonium, Perceptrons and so on, (ie, the pre-Robertswork). It should discuss the main ideas of modern visionwork as a leadin to the more specific articles, for

impIt fcht use of hypothesis, model, or expectationdriven strategies. It should also discuss the way inwhich the focus of the field flip-flops from front endconsiderations to higher level considerations with. I T 'c ,

Polyhedral or Blocks World Vision

An overview article should include the majorideas In this work together with briefsummaries of the work of the major Investigators,In addition, separate articles should be writtenon the work of {holt listed below.

Overylew NEED [73o-rts, Huffman and Clowes, Kelley,

Shlral and others listed below)

Guzman done (23

Falk NEED [53

Waltz NEED (73This article should contain more generalmaterial on constraint satisfaction, drawnpossibly from Montenarl and Fikes

This exhausts my list. Please add others or delete someOf mine 1 f appropriate.It has been suggested [Bollesl that the most instructivemethod of writing these articles would be to providesimple examples of the problems attacked by the variousprograms*

Ictn" .n"1 v" . "Overview NEED [93

This article should describe or point to detailedltrptofl.fi used, and the present state of theart .following articles should be written or modified to

describe the specialized tools of scene analysis.See Duda and Hart,

Template Patching NEED [53

' 10*36

"

" r.E.

F.

"

THU 15 APR 76 Psge 11.1(a non-mathematical description)

Edge Detection

AIH, OUTLINE;!

done (43

done t73Homogeneous CoordinatesThis article should be modified to Includethe general questions of the perspectivetransformation, camera calibration, and so on.

l ine Descr ipt lon

Noise Removal

done [43

done [43

done [43

done [33

NEED [43

NEED [43

Shape Description

Region Crowing

Contour Fol lowing

Spatial filtering

Front End Particulars NEED (63This article should contain some description of the■net hods and effects of compression and quantizationfor example.

Syntactic Methods NEED [53

NEED (6)Descriptive M ethodsSe« Duda md rlort# and Winston

Robot and Industrial Vision Systems

Overview ar d State of the Art NEED [93

NEED (83Hardware

Pattern Recognition

It's not fl.tfr Just where this discussion should go, orwhat level of detail is required.Overview done (81

This article needs to be refocussed and cleaned up

Statistical Methods and Applications

r>e"crlot Ive Methods and Applications

NEED [93

NEED [83

v 1 seel l aneous

Multlsensery Images

Percept rons

NEED [73

NEED t63

V/Vl_ _>«*€*** vA O-^wfiE-i-

4

"

"

"

Page 12 AIH, OUTLINE;!

IX, SPEECH UNDERSTANDING SYSTEMS

Overview (include a mention of ac, proc) done [33Integration of Multiple Sources

o* *ncw!edge NEED [9]For sxsmpls the blackboard of the HEARSAY II system

HEARSAY I done [43

HEARSAY II done [53

BPEECHLXB done [23

SDC-SRI System NEED (73

DRAGON done [63Tim Baker's originel system plus Speedy-Dragon byBruce Lowerre. This article is a little harder thanthe other system articles because the methods usedmay be unfamiliar to some.

10136 THU 15 APR 76

4 10:36 THU 15 APR 76 Page 13 AIH, OUTLINE;!

"

"

X, ROBOTICSmOverview NEED (93

This article should discuss the central issues anddifficulties of the field, its history, and thepresent state of the art.

Robot Planning and Problem Solving NEED [83For example, strips and ABSTRIPS, This articlecould Of quite general depending on the point ofviaw taken,

* rns NEED [8]Explain the difficulties of control at the bottomlevel, system Integration, obstacle avoidanceond so on. Also note the problems with integrationOf mv! t i -sensory data, for example vision andtouch feedback.

Present Day Industrial Robots NEED [73

NEED [63Robotics Programming LanguagesFor txtmplt WAVE, and AL

(a short article)

10:36 THU 15 APR 76 Page 14 AIH. OUTLINE;!

"

"

"

XI, Appl i cat ions -OF Ai '&ai*^v s^onsAn overview article. What are the attributes NEED [83of a suitable domain ? Custom crafting -theory vs. actual us?, Csoo EAF; ?25 notes, 1972)

A. Chemistry

1. Mass spectrometry (DENDRAL, CONGEN, met a-dendral ) done [632, Organic Synthesis

Overview NEED (81Corey, Gelernter, and SridharanSummarize work of Mpke,

S. Medicine1, MYCIN done (13

CASNETCKul ikowski), NEEDC732, Summarize DTALOGCPopI e) ,Pauker's MH work, and the Genetics counsellingprograms

C. Psychology and PsychiatryProtocol Analysis (Waterman and Newell) NEED 163

"". Mitt, systems

1. REt MEED [4]NEED [632. MACSYMA (mention SAINT)

E« Business and Management Science Applications

1. Assembly line balancing (Tonge)2, Flectric power distribution systems

MEED [53NEED C53

(MI)

F. Mi seel l aneous

1. LUNAR2, Education

NEED [5]NEED C73

Papert, or more ?3, SRI computer-based consultation4. RAND--RITA production rule system for

NEED [63NEED [53

intelligent interface software

I. Miscellaneous

Overview of music composition and aesthetics done [73

*r*t.oATiO*-^(^ P*^ *»\r %t-CWW>t fob vilM