Program & Schedule · involved in integrating expert system technology in commercial environments...

T H E E I G H T H

N A T I O N A L

C O N F E R E N C E

O N A R T I F I C I A L

I N T E L L I G E N C E

AAAI–90

Program

& Schedule

Contents:

• AI On-Line / A–2

• General Information / A–8

• Conference at a Glance / A–12

• Program / A–14

• Index / A–25

• Maps / A 11, A–29

Sponsored by the

American Association for

Artificial Intelligence

July 29, 1990 – August 3, 1990

• How will it be different/same

(5) Assessment of critical success fac-tors:• How they established criteria• Whether the system met the crite-

ria or not• Approaches for measuring future

system success

PANELISTS

■ Larry Tieman, Managing Director ofData Management/Knowledge Engi-neering Systems, American Airlines

American Airlines has deployedmany significant expert system appli-cations on a variety of hardware plat-forms from mainframes to work-stations to personal computers. Twosignificant expert system applicationsand their impact on American Air-lines’ business and organizations willbe presented.

■ Don McKay, Director, Logistic Sys-tems, American President Companies

Logistics management application.Mainframe expert system responsiblefor assigning vessel stowage locationsto move cargo between Asia andNorth America.

■ Roger Jambor, Director, BusinessInformation Group, Dun & Brad-street

The Credit Clearing House ExpertSystem is responsible for providingon-line credit analysis and recom-mendations to manufacturers, whole-salers, and importers on businesses inthe apparel industry.

■ John Woods, Manager of AI/ExpertSystem Technology DevelopmentGroup, Ford Motor Company

Ford has achieved several expert sys-tem application successes. One inparticular, ESCAPE (Expert System forClaims Authorization and Processing),is a mainframe expert system that isused in the validation process forincoming warranty claims andreplaces a major portion of an existingCOBOL program. It is an excellentexample of expert system technologysolving a major business problem.

■ Jim Euchner, Supervisor of ExpertSystem, Nynex Science & Technology

Nynex has deployed an expert systemin over 40 offices throughout NewEngland and its other service areas tohandle customer telephone repairproblems. It is having a dramaticimpact on reducing operational costsas well as improving customer service.

■ Rich Levine, President, RSL BusinessSolutions (Independent Consultant),Sun America Financial

Sun America Financial has deployedan Appointment Expert System toprovide expert advice for the corpo-rate and individual agent appoint-ment process in compliance withstate and federal insurance licensingregulations. This expert system is acritical component in a new process-ing automation system that inte-grates image processing, PC net-works, workflow management, andinstantaneous MIS tracking to rede-fine a business operation.

Why and How Corporate America IsIntegrating AI into Data ProcessingWednesday, August 1, Room 210, HynesConvention Center, 9:00 AM – 11:00 AM

Session organized by Jack Rahaim,Manager of AI Marketing, DigitalEquipment Corporation

■ Moderator: Don Mick, Director ofBusiness Opportunities, Aetna Life &Casualty

This session will focus on the fol-lowing three areas:• Applications • Architecture • Organizational Processes

PANELISTS

■ Ted Smith, US West

Ted Smith will discuss the topic ofapplications and emphasize theirorganization’s use of the technology.He will also discuss what applications

A–2 AI ON-LINE

AI On-LineSix Companies Reveal Results of Commercial Expert System SuccessesTuesday, July 31, Room 210, Hynes Con-vention Center, 2:00 PM – 4:00 PM

Session organized by Chuck Williams,Founder and Chief Technical Officer,Inference Corporation and FrankAngrisani, President, Sun AmericaFinancial

■ Moderator: Frank Angrisani, Presi-dent, Sun America Financial

This panel will review the dynamicsinvolved in integrating expert systemtechnology in commercialenvironments to solve major businessproblems. Panelists will share theirexperiences (the good and the bad) inusing expert system technology toimprove their organizations opera-tions. The panel will also take a broad-er look at expert system technology’spotential in future systems. Executivesrepresenting major corporations inthe telecommunications, transporta-tion, finance, insurance, and manu-facturing industries will participate.

Each panelist will present:

(1) a brief description of the expertsystem application(s) in production:• Business problems it solved• Technical approach

(2) What benefits were achieved:• Financial• Strategic• Development/maintenance pro-

ductivity• Quality

(3) Lessons learned: • Management• Integration• Development methodology• Organizational impact

(4) Plans for expert system technolo-gy in the future:• Applications types• Benefits sought• Methodology/technology to be

employed

choices were made, why they weremade, and how they were made.

■ Rosemary Baer, Vice President,Chemical Bank New York, ChemicalBank

Rosemary V. Baer will discuss Chemi-cal Bank's “retail architecture” andthe enabling characteristics whichpermit the continued integration ofexisting and emerging expert andconventional technologies. The dis-cussion will include the businessgoals and objectives which fosteredthe design of this highly distributed,service-oriented environment.

■ Paul Deschamps, Project Manager,Aetna Life and Casualty

Paul Deschamps will emphasize orga-nizational processes. He will talkabout their organization’s approachto learning, transferring and nurtur-ing expert systems throughout semi-autonomous divisions. Mr. Des-champs will also discuss the lessonslearned in applying the technologyand the current status vis a vis origi-nal goals of this effort.

■ Nancy Faucheux, Information Sys-tems Analyst, Texaco Inc.

Nancy Faucheux from Texaco Inc.will describe on-line applications ofreal-time advisory systems in petro-leum and petrochemical operations.

What Users Are Buying Today in AI –– and Why: Reports from ManagementWednesday, August 1, Room 210, HynesConvention Center, 1:00 PM – 3:00 PM

Session organized by Keith Scovell,Manager, and Tim Holcomb, SeniorConsultant, Andersen Consulting

■ Moderator: Roger Schelm, Vice Pres-ident, Applied Research, CIGNA

Mr. Schelm is Vice President,Applied Research/Expert Systems, of

AAAI ‘90 A–3

AI On-Line Schedule

Tuesday, July 31

Six Companies Reveal Results of Expert Systems Successes2:00 PM – 4:00 PM, Room 210, Hynes Convention Center

Organized by Chuck Williams, Inference Corporation and Frank Angrisani, Sun America Financial

Moderator: Frank Angrisani, President, Sun America Financial

Panelists: Larry Tieman, American Airlines; Don McKay, American President Companies; Roger Jam-bor, Dun & Bradstreet; John Woods, Ford Motor; Jim Euchner, Nynex; Rich Levine, Sun America

Wednesday, August 1

Why and How Corporate America Is Integrating AI into Data Processing9:00 AM – 11:00 AM, Room 210, Hynes Convention Center

Organized by Jack Rahaim, Digital Equipment Corporation

Moderator: Don Mick, Director of Business Opportunities, Aetna Life & Casualty

Panelists: Ted Smith, US West; Rosemary Baer, Chemical Bank; Paul Deschamps, Aetna Life andCasualty; Nancy Faucheux, Texaco Inc.

What Users are Buying Today in AI and Why: Reports from Management1:00 PM – 3:00 PM, Room 210, Hynes Convention Center

Organized by Keith Scovell and Tim Holcomb, Andersen Consulting

Moderator: Roger Shelm, Vice President Applied Research, CIGNA

Panelists: Mary Dunn, Mutual of New York; Bill Corbin, Owens Corning; Dave Wise, Frito-Lay; MikeRogalski, Sears Merchandise Group; Troy A. Heindel, NASA–Johnson Space Center; Al Brown, Lock-heed Missiles & Space Company

Developing Real-Time On-Line Applications3:15 PM – 5:15 PM, Room 210, Hynes Convention Center.

Organized by Didi Murnane, Gensym Corporation

Moderator: Karlene A. Kosanovich, Senior Consulting Engineer, E.I. DuPont de Nemours and Company.

Panelists: Dorothy Yu, Coopers & Lybrand; Karl-Erik Arzen, Lund Institute of Technology; GregoryM. O’Connor, Massachusetts Institute of Technology; Paul D. Schoen, Rockwell International; Jean-Pierre Aubert, Alcatel ISR France

Thursday, August 2

Differentiating Expert Systems Products9:00 AM – 11:00 AM Room 210, Hynes Convention Center

Organized by Larry Harris, AICorp and E. Robert Keller, Renaissance International

Moderator: E. Robert Keller, President, Renaissance International

Panelists: Larry Harris, AICorp; Harry C. Reinstein, Aion Corporation; Randy Raynor, Inference Cor-poration; Alain Rappaport, Neuron Data; Robert Moore, Gensym

AI Versus Conventional Data Processing2:00 PM – 4:00 PM, Room 210, Hynes Convention Center

Organized by Keith Scovell and Tim Holcomb, Andersen Consulting

Moderator: Bruce B. Johnson, Director of Research and Technology Services, Andersen Consulting

Panelists: Bradley C. Billetdeaux, Exxon Company USA; Stanley Wozniak, MCI; Dave Dean, East-man Kodak; Glen Galen, Burlington Northern Railroad Information Systems; Ed Mahler, E. I.DuPont de Nemours and Company

the Systems Division of CIGNA Cor-poration. He was instrumental in thelaunching of Expert Systems: Plan-ning/Implementation/Integration, aquarterly journal published by Auer-bach, a Division of Warren, Gorham& Lamont, and continues as itsfounding Consulting Editor. Mr.Schelm has directed the technologytransfer of expert systems intoCIGNA Corporation since 1983.CIGNA Corporation is one of theleading users of expert systems in theinsurance industry.

This panel discussion surroundshow and why businesses are apply-ing AI technology to specific areaswithin their organization, their basisfor selecting and applying AI to eachdomain, the process of identifyingopportunity areas and building abusiness case, the effectiveness oftechnology transfer within the orga-nization, and the benefits realized asa result. Each issue will be supportedby examples of how and why knowl-edge-based systems (KBS) applica-tions and decision-support systemsare being applied within the corpo-rate community. In order to facilitatethis discussion, the topic will beginby defining and describing AI from acommercial view, rather than an aca-demic (or theoretical) perspective,and conclude with a review of cer-tain key benefits realized from thedevelopment and delivery of KBSs.

Each panelist will present:

(1) a brief description of major KBSapplications in production:• Business problem(s) solved

(the need)• Development considerations /

approach

(2) The key benefits realized:• Improved workforce productivity• Improved product/service quality• Reduced corporate expenditures• Reduced system development /

maintenance costs

(3) Brief description of businesscase(s) developed:• Success criteria/critical factors• Key project approach• Measuring KBS success

(4) Lessons learned:• Development approach• Organizational change issues• Knowledge automated

PANELISTS

■ Mary Dunn, Corporate KnowledgeEngineer, Mutual of New York

Mutual of New York (MONY) tookadvantage of KBS technology relative-ly early, implementing an underwrit-ing system shell in 1987 in whatmight be called “first generation”expert-system shell software. Thisfirst project was a major undertaking,not cautious experimentation, and amajor success. Where do you go fromthere, particularly in a climate ofdownsizing and cost cutting? Threeobjectives serve our goal of seeingKBS technology become a power toolon the business systems workbench.These objectives are:• Integrate the technology into the

mainstream of systems develop-ment and operations

• Integrate knowledge engineeringinto business systems support areas

• Integrate understanding of thetechnology’s applicability intobusiness areas (management andanalysts)Ms. Dunn will describe the steps

taken in this “second generation”KBS phase and the lessons learnedalong the way.

■ Bill Corbin, Systems Engineer,Owens Corning

Owens Corning is the leading manu-facturer in the U.S. for a wide varietyof fiberglass and flaked glass productsranging from the batting used toinsulate homes to fiberglass-rein-forced plastic parts for refrigerators.The company uses a decision-supportsystem that monitors a large convey-or belt that produces continuousmats of fiberglass that must meet cer-

tain quality standards, such as tensilestrength and uniform size and thick-ness. Operating around the clock, thebelt is subject to numerous mechani-cal failures that could affect thesequality standards. Belt speed, oventemperatures, and settings for themelters must be constantly moni-tored and fine-tuned in order for theprocess to run smoothly.

The decision-support system comesinto play when there is a problem.For instance, if a problem were toarise regarding the strength of themat, the expert system would takethe operator through an interviewprocess to help determine what iscausing the problem. The applicationruns on the Macintosh with a hyper-text front-end, and shows the opera-tor graphically exactly where to lookin the machinery to find the possiblesource of the problem.

■ Dave Wise, Senior Systems Analyst,Frito-Lay

Frito-Lay is developing a knowledge-based system which can be used tomake decisions by trolling throughdata gathered from over 10,000hand-held computers, purchasedscanner data and other internal datasources. Discussion will be centeredaround the development of the busi-ness case, criteria used to justify pro-jects and problems encountered todate. Future directions include theuse of analytical workbenches.

■ Mike Rogalski, Software Specialist,Sears Merchandise Group

Mr. Rogalski directs efforts to investi-gate and deploy AI technologies andknowledge-based systems for SearsMerchandise Group. His talk will out-line some of the areas in whichknowledge-based systems are beingdeveloped within the company, withattention paid to benefits –– expectedand unexpected –– for the businessand for the software developmentorganization.

A–4 AI ON-LINE

■ Troy A. Heindel, Project Manager,Real-Time Data System, NASA–John-son Space Center

NASA is taking full advantage of AItechnology, developing real-timeexpert systems in the Johnson SpaceCenter Mission Control Center toassist space-shuttle flight controllers.NASA is using KBSs to address threebasic goals: (1) capture of corporateknowledge of veteran flight con-trollers, (2) enhancement of flightcontroller tools, and (3) reduction inthe time required to train new per-sonnel. Mr. Heindel has experiencein the design and implementation ofKBSs, which support various problemdomains including flight simulation,reaction control, and fault diagnosis.

■ Al Brown, Project Manager, Lock-heed Missiles & Space Co.

Mr. Brown will use his experiencewith the Advanced Solid RocketMotor project to demonstrate somepositive steps to be taken in obtain-ing management commitment for anAI project. At Lockheed, this includ-ed putting something that was notproven and had not been done in aproposal to NASA. Mr. Brown willalso discuss some of the things thatshould not be done when trying tosell management on AI. On theASRM project, Lockheed had to con-vince all contractors involved tomove AI into the mainstream of theproject.

Developing Real-TimeOn-Line Applications

Wednesday, August 1, Room 210, HynesConvention Center, 3:15 PM––5:15 PM

■ Session organized by Didi Murnane,Gensym Corporation

■ Moderator: Karlene A. Kosanovich,Senior Consulting Engineer, E.I.DuPont de Nemours and Company

Karlene A. Kosanovich is a senior

consulting engineer in the ArtificialIntelligence in Process ControlGroup at DuPont. Ms. Kosanovich isactively involved in the developmentand implementation of real-timeexpert systems aimed at improvingall aspects of process control.

The aim of this panel is to providea forum in which the special needs ofreal-time on-line applications will beshared. Each panelist will describetheir project from conceptual designto implementation. Key issues to bediscussed will be lessons learned, pro-ject development and implementa-tion barriers, and real business bene-fits of the technology. Each panelistwill present:

(1) A definition of their applicationdomain

(2) Application issues addressed

(3) Application implementation

(4) Key issues:• Problems solved by application• By-product benefits of application• In retrospect, issues that would be

dealt with differently

PANELISTS

■ Dorothy Yu, Manager AdvancedTechnology Group, Coopers &Lybrand

The Coopers & Lybrand AdvancedTechnology Group has been usingknowledge-based process modelingand simulation to improve back-office operations in the financial ser-vices industry. The C & L ProcessSimulator developed using the G2Real-Time Expert System shell, isused to model, analyze, and redesignvalue-added activities and resourceutilization as a means to facilitateand improve on-line processing.

■ Karl-Erik Arzen, Department ofAutomatic Control, Lund Institute ofTechnology, Sweden

The Department of Automatic Con-trol at the Lund Institute of Technol-ogy has been working in conjunctionwith Asea Brown Boveri AB, andSattControl AB on the research pro-

ject, knowledge-based real-time con-trol systems. Karl-Erik Arzen is theprincipal person in the AutomaticControl Department on this project.Mr. Arzen’s background and experi-ence is in expert control where a real-time, blackboard-based expert-systemshell has been developed and imple-mented. The system currently on-lineis used as a part of a feedback loopcontaining general control knowl-edge and heuristics on control looptuning, monitoring, and adaption.

■ Gregory M. O’Connor, PostdoctoralFellow, Biotechnology Process Engi-neering Center (BPEC), Mas-sachusetts Institute of Technology

Mr. O’Connor’s work at MIT’s BPECfocuses on the use of automation tech-nologies to improve the consistencyand quality of bioprocess operationsand to speed process development. Hehas built on-line expert systems toensure product quality and superviseprocess operations. The MIT BPEC isan Engineering Center of Excellencesponsored by the National ScienceFoundation, to foster closer university-industry relations. The BPEC has anindustrial consortium with more than60 member companies, many ofwhich are interested in advanced com-puter control of bioprocesses.

■ Paul D. Schoen, Project ManagerAerospace Simulation and SystemsTest Center, Rockwell International

The Aerospace Simulation and Sys-tems Test Center in Downey, Califor-nia builds automatic test equipmentand simulation systems like NASP(National Aerospace Plane). Theyhave produced shuttle flight-simula-tion systems and are the back-updata-acquisition center for Houston’sJohnson Space Center to receive in-flight shuttle data. The center isRockwell’s Center of Excellence forall simulation activities. Paul Schoenis the project manager in charge ofdeveloping ground support equip-ment that is utilized for data acquisi-tion and reduction including the useof expert systems.

AAAI ‘90 A–5

■ Jean-Pierre Aubert, Division Manag-er of C3I Systems, Alcatel ISR France

Jean-Pierre Aubert is Division Manag-er in charge of C3I systems at AlcatelISR, Paris, France. He will presentSICOCS, a communications com-mand and control system for fight-ing forest fires. The systems locateteams of fire fighters, allocatesresources appropriately, and main-tains historical documentation ofeach mission. The entire system iswritten in SPOKE object-orientedprogramming language and inter-faces to DOG NAPPING and Oracle.

Differentiating Expert System ProductsThursday, August 2, Room 210, HynesConvention Center, 9:00 am – 11:00 am

■ Session organized by Larry Harris,Founder, AICorp and E. Robert Keller,President, Renaissance InternationalCorporation

■ Moderator: E. Robert Keller, Presi-dent, Renaissance International Corp.

In February 1983, Mr. Keller formedRenaissance International. Its pur-pose is to provide support and train-ing services, education, and productsin the AI field .

Immediately prior to forming Ren-aissance, Mr. Keller served for threeyears in an executive capacity as Direc-tor of Customer Support and Relationsat Artificial Intelligence Corporation.In that position, he was responsiblefor directing work in the areas of prod-uct planning and evaluation as well aspersonnel, public, and customer rela-tions. In addition, he has spent severalyears as a self-employed systems con-sultant and teacher of both AI andstructured system development.

Segment 1: Five minute introductionby the moderator outlining the issues,goals, and methods of the panel.

Segment 2: Ten minute presentationsby each of the five panelists in which

they differentiate their products fromthe rest of the market, specificallyaddressing the following three areas:

(1) Major design goals of the product

(2) Unique functional and architec-tural features that differentiate theproduct from others

(3) Two production applications thatbest illustrate the product’s capabili-ties. Only applications that are inproduction use by customers areacceptable. Include the following:• Nature of the application• Specific benefits to the customer• Typical number of users per day,

and maximum concurrent users• Size and nature of databasesinvolved

Segment 3: 30 – 60 minutes of Q&A.Questions are from the audience, aswell as from the moderator.

PANELISTS

■ Larry R. Harris, Founder, AICorp

Mr. Larry R. Harris is the founder ofAICorp. He is the author of the INTEL-LECT natural language informationsystem, and the chief architect ofKEMS, AICorp’s Knowledge Base Man-agement System for production appli-cations in the IBM mainframe envi-ronment. Mr. Harris is an internation-ally recognized authority on allaspects of AI technology, especiallywith regard to natural language andexpert systems. During his tenure asProfessor of Computer Science at Dart-mouth College, he wrote the INTEL-LECT system. While at Dartmouth, hewas also a visiting professor at theMassachusetts Institute of Technolo-gy’s AI Laboratory. Mr. Harris is theauthor of the Bantam Book, ArtificialIntelligence Enters the Marketplace.

■ Harry C. Reinstein, Aion Corporation

Harry C. Reinstein is currently theChairman and Chief Executive Offi-cer of Aion Corporation. In additionto being one of Aion’s founders, hewas president from its inception in1984, through 1988. Prior to Aion, Mr.Reinstein was with the IBM Corpora-

tion for 23 years. Throughout his ca-reer he has been directly involved inproduct development and customerinteraction relating to production-cen-tered data processing. Specific projectshave included operating systems,database management, applicationdevelopment, and expert systems.

■ Randy Raynor, Senior Vice Presi-dent, Inference Corporation

Mr. Raynor joined Inference Corpora-tion in July 1987. In the position ofSenior Vice President, he is responsi-ble for the development and supportof all Inference products. Mr. Raynorcame to Inference with over twelveyears experience in management ofmajor software organizations. Prior tojoining Inference, Mr. Raynor wasVice President, Product Developmentfor Computer Associates (formerlyUCCEL), where he managed the orga-nization responsible for developingproducts for the company’s businessapplication division. Prior to Com-puter Associates, he was with TexasInstruments where he held variousmanagement positions in develop-ment of computer-aided softwareengineering tools (IEF) as well as busi-ness and engineering applications.

■ Alain Rappaport, President, ChiefScientist, Neuron Data

Mr. Rappaport is responsible for thedesign of NEXPERTOBJECT reasoningtechnology. Prior to cofounding Neu-ron Data, he was a researcher in AI andmachine learning at the Robotics Insti-tute at Carnegie-Mellon University.

■ Robert Moore, President, Gensym

Mr. Moore, along with five othermembers of Gensym, founded thecompany in September 1986. Thecompany was founded to developand market a new generation ofexpert system products—productsthat would provide capabilities forreal-time data for time-critical appli-cations. Mr. Moore has been theauthor of several articles on real-timeexpert systems and has presentednumerous papers at AAAI and ISA.

A–6 AI ON-LINE

AI Versus Conventional Data ProcessingThursday, August 2, Room 210, HynesConvention Center, 2:00 PM – 4:00 PM

Session organized by Keith Scovell,Manager, and Tim Holcomb, SeniorConsultant, Andersen Consulting

■ Moderator: Bruce B. Johnson, Direc-tor of Research and Technology Ser-vices, Andersen Consulting

Mr. Johnson is Director of the Centerfor Strategic Technology and Research.He founded Andersen Consulting’s AIpractice in 1982. Mr. Johnson is ac-tively involved in the research and de-velopment of knowledge-based appli-cations, distributed decision-makingplatforms, and integrated human-computer information systems.

This panel will provide an intelli-gent comparison between AI andconventional data-processing tech-nologies as they relate to providingsolutions to major business prob-lems. During this session, each pan-elist will share their experiences andviews on the two approaches andsupport their beliefs with exampleKBS applications. The session willcover such topics as assessing busi-ness opportunities, building a casefor AI/conventional DP, and key ben-efits expected. Each issue will be sup-ported by examples of how and whyAI/conventional DP is applied to cor-porate business problems. Each pan-elist will present:

(1) Brief description of major KBS sys-tems in use:• Business problem(s) solved• Reason for the selection of an AI

solution

(2) Comparison of AI versus conven-tional data-processing approaches:• Automation opportunities• Development methodologies• Examples of current conventions /

systems which are solid candidatesfor AI applications.

(3) Key benefits achieved:

• Reduction in system maintenancecosts

• Reduction in development time-frame

• Improved operating efficiencies

(4) Architectures for integration

PANELISTS

■ Bradley C. Billetdeaux, Senior Sys-tems Analyst, Information and Com-munications Systems Department,Exxon Company, USA

Mr. Billetdeaux’s team provides con-sulting and development services forICBS and operations research applica-tions. Integration and maintainabili-ty are issues when consideringknowledge-based systems (KBS) forapplications development. At ExxonCompany, USA, the experience incor-porating three KBS modules within amaterial inventory system providedvaluable insight on how to addressthese issues. Procedures for cost esti-mating, staffing, documentation, andmainframe performance help toensure that the potential benefitsfrom using KBS are captured.

■ Stanley Wozniak, Manager, Knowl-edge Based Systems Department, MCICommunications Corporation

Mr. Wozniak‘s organization is respon-sible for applying advanced tech-nologies to support MCI’s strategicbusiness objectives. Currently theKBS Department focuses its efforts onbuilding intelligent front-ends tocomplex database applications in thesales and marketing, and the order-entry areas of MCI’s business. Otherprojects include integrated networkalarm management, strategic busi-ness modeling, and intelligent andintegrated application environments.

Mr. Wozniak will discuss how hisdepartment is designing and integrat-ing an AI-based software develop-ment life-cycle into a traditional one,MCI’s corporate systems develop-ment methodology (CSDM).

■ Dave Dean, Systems Engineer, AILab Coordinator, Eastman Kodak

The Eastman Kodak Company is utiliz-

ing a knowledge-based systemapproach to solving problems inmany areas. Virtually hundreds ofdecision-tree based expert systems arein place, used by technicians to diag-nose manufacturing machine relatedproblems. Expert systems have beeninstalled where conventionalapproaches were unsuccessful, such asin a warehouse picking operation, orinstalled in addition to a convention-al approach, such as a product selec-tion advisor that takes advantage ofan existing mainframe database sys-tem. Kodak has experience in expertsystems on a wide range of hardware /software platforms, as well as experi-ence in other advanced computingtechnologies such as neural networks,object-oriented systems, and intelli-gent human/machine interfaces.

■ Glen Galen, Manager—AI Systems,Burlington Northern Railroad Infor-mation Systems

Burlington Northern Railroad has beeninvolved in the development of artifi-cial intelligence-based systems sinceApril 1986, when Mr. Galen joined thecorporation. A number of expert sys-tems are now in use, running on bothpersonal computers and engineeringworkstations. These fielded applica-tions are in marketing, operations,accounting, and human resources.Burlington Northern also uses themainframe-based AI tools, KBMS, andINTELLECT, and has developed proto-type applications using that software.

■ Ed Mahler, Manager, Decision Sup-port Systems and Artificial Intelli-gence, E. I. DuPont De Nemours & Co.

Mr. Mahler’s organization is responsi-ble for evaluating and applying deci-sion-support technologies to supportE. I. DuPont’s corporate vision. E. I.DuPont has successfully fielded anumber of AI/KBSs in a variety offields. His discussion will centeraround business decision flow mod-els: the role of transaction systems,decision-support systems, knowl-edge-based systems, and models inthe decision-making process.

AAAI ‘90 A–7

AdmissionEach conference attendee will receivea name badge upon registration. Thisbadge is required for admittance tothe technical, tutorial, exhibit, andworkshop programs. Admittance tothe tutorials is by TICKET ONLY.

SMOKING, DRINKING, AND EAT-ING ARE NOT ALLOWED IN ANY OFTHE TECHNICAL, TUTORIAL, OREXHIBIT SESSIONS.

Baggage ClaimThere will be a baggage claim/checkarea at the Plaza level, Main Lobby inthe Hynes Convention Center oper-ated by ARA Leisure Services. Neitherthe AAAI nor the Hynes ConventionCenter accept liability for the loss ortheft of any suitcase, briefcase, orother personal belongings brought tothe site of AAAI-90.

Banking/Currency ExchangeThe closest full service bank to theHynes Convention Center is theBank of Boston located on PrudentialPlaza; the phone number is (617)267-5261. Deak International pro-vides currency exchange services attheir location, 800 Boylston Street inthe Prudential Plaza. Their hours areMonday—Friday 8:30 AM – 4:30 PM;Saturday 10:00 AM – 5:00 PM; Sunday 10:00 AM – 2:00 PM.

Child Care Personal Touch provides child-careservices in your hotel room. Cost is$12.00 per hour for 1-3 children witha minimum of four (4) hours.Arrangements can be made to havethe charges added to your hotel bill.Contact (617) 423-2556 for furtherdetails.

Coffee Breaks Coffee breaks for the Technical Ses-sions will be held in the Hynes Con-vention Center, Boylston Street Hall,Level 2 and 3.

Conference Program ScheduleThe program schedule begins on pageA-14.

CopiersThe closest copying centers to theHynes Convention Center are SirSpeedy at 827 Boylston Street, (617)267-9711 (Hours: M – F, 8:30 AM –5:00 PM; Sat., 10:00 AM – 5:00 PM), andCopy Cop at 815 Boylston Street,(617) 267-9267 (Hours: M – F 7:30 AM – 11:00 PM; Sat. 8:30 AM –6:00 PM; Sun 12:00 PM – 8:00 PM).

Digital Equipment VideoAn educational video about innova-tion in AI aimed at the secondaryschool level will be demonstrated inRoom 312 of the Hynes ConventionCenter. It will be shown continuouslyWednesday and Thursday, August 1-2from 8:30 AM – 6:00 PM.

Exhibit EntranceConference attendees must be wear-ing their conference registration orexhibitor badge to enter the Exhibi-tion or AI On-Line. Vendor-issuedguest passes must be redeemed at theregistration desk in Hall B. Furtherinformation regarding access to theExhibition can be obtained from theExhibitor Information Desk in frontof Exhibition Hall C.

Exhibit ProgramAn important service to conferenceattendees is the Exhibit Program, tobe held Tuesday, July 31 throughThursday, August 2. Hardware andsoftware manufacturers, publishers,universities, and nonprofit organiza-tions involved in artificial intelli-gence will display and demonstratetheir current products, services, orresearch.

Once again AAAI has given compli-mentary booths to university andnonprofit research organizations todemonstrate current AI research. TheAAAI would like to thank the majorcomputer equipment suppliers for

donating equipment and technicalsupport toward this demonstrationprogram.

LOCATIONHynes Convention Center, Hall C & D

EXHIBIT HOURSTuesday: 12:00 PM – 7:00 PM

Wednesday: 10:00 PM – 6:00 PM

Thursday: 10:00 PM – 5:00 PM

ExhibitorsThe AI Review has a complete list ofexhibitors, booth locations, and out-line of products, services or researchefforts. In addition, there are shortarticles submitted by exhibitors dis-cussing technical aspects of theirproducts or services. Extra copies ofthis review are available at theExhibitor Information Desk.

Food OutletsThere is a cafeteria in the Hynes Con-vention Center, on the Plaza level,main corridor. It will be open duringthe conference.

Handicapped FacilitiesThe Hynes Convention Center andthe Marriott Copley Place Hotel bothhave handicapped facilities. TheHynes Convention Center has eleva-tors located in the North and SouthRotundas connecting all three floors.

Information DeskAn information desk will be staffedduring registration hours, Sundaythrough Friday, July 29 – August 3.

List of AttendeesA list of preregistered attendees of the conference will be available for reviewat the Registrar’s Desk in the HynesConvention Center, Hall B. Attendeelists will NOT be distributed.

Message CenterA message desk will be manned inthe Hynes Convention Center, HallB, during registration hours, and mes-

A–8 GENERAL INFORMATION

General Information

sage terminals for sending andretrieving messages are located inHalls B, C, and D. The telephonenumber for leaving messages only is(617) 954-3333.

THERE IS NO MEANS OF PAGINGCONFERENCE ATTENDEES. Allphone messages can be retrievedfrom any message terminal in Hall Bduring registration hours or at themessage terminals in the Exhibit Hallduring exhibit hours. There will beno access to message terminals afterhours. We suggest that hotel phonenumbers be used as primary contactpoints.

ParkingPaid lots are available under the Pru-dential Towers adjacent to the HynesConvention Center for $14.00 perday and on Dalton Street across fromthe Sheraton at the Cheri Lot for$10.00 per day.

Press All members of the media arerequested to register in the PressRoom on the Plaza Level of theHynes Convention Center, Room108. Press badges will only be issuedto individuals with approved creden-tials. The Press Room will be open foradvance registration on Sunday, July29 at 12:00 PM. During the confer-ence the Press Room will be openduring the following hours:

Sunday, July 29: 12:00 PM—5:00 PM

Monday, July 30: 8:00 AM—5:00 PM

Tuesday, July 31: 8:00 AM—5:00 PM

Wednesday, August 1: 8:00 AM—5:00 PM

Thursday, August 2: 8:00 AM—5:00 PM

Friday, August 3: 8:00 AM—11:00 AM

A representative from the AAAIwill be on duty during press-roomhours to assist the members of thepress and media.

Press ConferenceA Press Lunch and Conference willbe held on Tuesday, July 31, at 12:00PM in Room 208 of the Hynes Con-vention Center.

ProceedingsEach registrant for the technical pro-gram will receive a ticket with theregistration materials for one copy ofthe conference proceedings. The tick-et may be redeemed at The MIT PressProceedings counter, located in theHynes Convention Center, Hall B,during registration hours, or at theAAAI Press / The MIT Press booth#306, located in Hall C, duringexhibit hours. Proceedings can alsobe redeemed by mailing the ticketwith your name and address to theMIT Press, 55 Hayward, Cambridge,MA 02142.

Extra proceedings may be pur-chased at the conference site at theabove locations. THURSDAY,AUGUST 2, WILL BE THE LAST DAYTO PURCHASE EXTRA COPIES OFTHE PROCEEDINGS AT THE CON-FERENCE SITE. Proceedings are alsoavailable at The MIT Press bookstorein Cambridge, MA.

Public TelephonesPay telephones are located on all lev-els of the Hynes Convention Centerand the Marriott Copley Place Hotel.

RecordingNo audio or video recording isallowed in the TUTORIALS. Tapes ofTECHNICAL sessions will be for salein Hall B, Hynes Convention Center.A representative from First Tape Inc.will be available to take your orderduring regular registration hours,beginning Tuesday, July 31. Orderforms are included with registrationmaterials. Tapes may also be orderedby mail from

First Tape Incorporated730 North LaSalle Street, Suite 200Chicago, Illinois 60610.

RegistrationConference registration will takeplace in the Hynes Convention Cen-ter, Hall B, beginning Sunday July 29.

Registration hours are:Sunday, July 29: 7:30 AM—6:00 PM

Monday, July 30: 7:30 AM—6:00 PM

Tuesday, July 31: 7:30 AM—6:00 PM

Wednesday, August: 1 7:30 AM—6:00 PM

Thursday, August 2: 7:30 AM—6:00 PM

Friday, August 3: 7:30 AM—11:30 AM

Only checks drawn on US banks,VISA, Mastercard, government pur-chase orders, and travelers’ checks forUS currency will be accepted. Wecannot accept foreign currency orchecks drawn on foreign banks.

Speakers Ready RoomsSpeakers Ready Rooms will be locatedon the second floor, Room 205 of theHynes Convention Center. This roomwill have audio-visual equipment toassist speakers with their prepara-tions. It is important that speakersutilize this room to organize theirmaterials. The room will be openduring the following hours: 8:00 AM –5:00 PM Sunday, July 29 throughThursday, August 2, and 8:00 AM –10:30 AM on Friday.

Invited Speakers are asked to cometo Room 111 one day prior to theirspeech. Room 111 is located on theFirst Level of the Hynes ConventionCenter and the audio-visual teamwill be there from 9:00 AM – 12:00 PM

every day.

T-Shirt SalesT-shirts will be for sale in Hall B, HynesConvention Center, during registrationhours. Supplies are limited.

AAAI ‘90 A–9

TransportationCustom Travel Consultants, theAAAI-90 travel agent, will be staffingthe Transportation Desk in the HynesConvention Center, Hall B, to assistwith changes in travel plans.

AirNorthwest Airlines and Trans WorldAirlines (TWA) have been selected asofficial co-carriers for AAAI-90. If youwish to change your reservation, youmay call Northwest directly at (800)328-1111, and TWA directly at (800) 325-4933. Please remember toprovide the AAAI-90 ConferenceCode number (NW) 17379, (TWA)B9912829 when you make yourarrangements.

Ground Transportation in BostonThe Massachusetts Bay Transporta-tion Authority, commonly called“The T,” services most of Boston.Hynes Conference Center can bereached by taking the Green Line tothe Prudential or Auditorium sta-tions. Cost at press time was $0.75for one direction.

Tutorial SyllabiAAAI-90 Tutorial syllabi will NOT beavailable after tutorials.

Volunteer RoomVolunteer Headquarters, located onthe Plaza Level of the Hynes Conven-tion Center in Room 107, will beopen Saturday, July 28.

Special Meetings

AAAI Annual Business MeetingThe Annual Business Meeting will beheld Wednesday, August 1, from12:45 PM – 1:15 PM in the Hynes Audi-torium.

AAAI Conference Committee MeetingThe AAAI Conference Committeebreakfast meeting will be held Tues-day, July 31, at 7:30 AM in Room 300(the Board Room), Level Three of theHynes Convention Center.

AAAI Executive Council MeetingThe Executive Council Meeting willbe held Monday, July 30, from 8:30PM – 5:00 PM, in Room 300 (the BoardRoom), Level Three of the HynesConvention Center.

AAAI Press Editorial Board MeetingThe AAAI Press Editorial Board willmeet Wednesday, August 1, at 1:20PM in Room 300 (the Board Room),Level Three of the Hynes ConventionCenter.

AAAI Publication Committee MeetingThe AAAI Publication Committeebreakfast meeting will be held Thurs-day, August 2, at 7:30 AM in Room300 (the Board Room), Level Three ofthe Hynes Convention Center.

AI in Business Subgroup MeetingThe AI in Business Subgroup Meetingwill be held Tuesday, July 31, from12:45 PM – 1:30 PM in Room 302 ofthe Hynes Convention Center.

AI in Law Subgroup MeetingThe AI in Law Subgroup Meeting willbe held Thursday, August 2, from12:45 PM – 1:30 PM in Room 302 ofthe Hynes Convention Center.

AI in Manufacturing Subgroup MeetingThe AI in Manufacturing SubgroupMeeting will be held Thursday,August 2, at 12:45 PM – 1:30 PM inRoom 304/6 of the Hynes Conven-tion Center.

AI in Medicine Subgroup MeetingThe AIM Subgroup Meeting will beheld Tuesday, July 31, from 12:45 PM

– 1:30 PM in Ballroom C of the HynesConvention Center.

AI Journal Editorial Board MeetingThe Artificial Intelligence Journal Edito-rial Board breakfast meeting will beheld Wednesday, August 1, at 7:00 AM

in the Vineyard Room, Level Four,Marriott Copley Place.

IJCAI Trustee's MeetingThe IJCAI Trustee’s Meeting will beheld Friday, August 2, from 8:00 AM –6:00 PM in the Vineyard Room, LevelFour, Marriott Copley Place.

Planning MeetingThe Knowledge Discovery in DatabasesPlanning Meeting will be held Thurs-day, August 2, in Hynes ConventionCenter, Room 208, from 10:00 AM –12:30 PM. (Gregory Pietetsky-Shapiro.)

Press Lunch and ConferenceA Press Lunch and Conference will beheld Tuesday, July 31, at 12:00 PM inRoom 208 of the Hynes ConventionCenter.

SIGART MeetingThe SIGART Meeting will be heldTuesday, July 31, from 12:45 PM –1:30 PM in Room 304/6 of the HynesConvention Center.

A–10 GENERAL INFORMATION

Alternative Activities

Tuesday—Open SuiteInitiative for Managing KnowledgeAssets, Tuesday, July 31, 2:00 PM –4:00 PM, Sheraton Boston Hotel, Suite507/508. Jennifer Oshin, CarnegieGroup

Wednesday—Open SuiteInitiative for Managing KnowledgeAssets, Wednesday, August 1, 2:00 PM

– 4:00 PM, Sheraton Boston Hotel,Suite 507/508. Jennifer Oshin,Carnegie Group

Wednesday—Franz ForumBirds of a Feather Session—Usergroup meeting/open discussion forthe Franz product line including Alle-gro Common LISP and future prod-ucts. Technical developers will be onhand. Wednesday, August 1, 3:00 PM

– 4:00 PM, Hynes Convention Center,Room 208

Thursday—Open SuiteInitiative for Managing KnowledgeAssets, Thursday, August 2, 2:00 PM –4:00 PM, Sheraton Boston Hotel, Suite507/508. Jennifer Oshin, CarnegieGroup

AAAI ‘90 A–11

Fourth Floor Map, Boston Marriot Copley-Place

Sunday, July 29Conference Registration7:30 AM – 6:00 PM, Hall B, Hynes

◆Tutorial Program2:00 PM – 6:00 PM

Tutorial SP1: Basic Theory ofNeural NetworksGeoffrey Hinton and Michael Jordan2:00 PM—6:00 PM, Ballroom B, Hynes

Tutorial SP2: Designing Natural Language InterfacesEdwin Addison and Paul Nelson2:00 PM—6:00 PM, Room 304/6, Hynes

Tutorial SP3: Model-Based DiagnosisWalter Hamscher and Peter Strauss2:00 PM—6:00 PM, Room 302, Hynes

Tutorial SP4: Constraint Reasoning: Theory and ApplicationsSanjay Mittal and Bernard Nadel2:00 PM—6:00 PM, Ballroom A, Hynes

Tutorial SP5: Distributed Artificial IntelligenceLes Gasser and Jeffrey Rosenschein2:00 PM—6:00 PM, Ballroom C, Hynes

Monday, July 30Conference Registration7:30 AM – 6:00 PM, Hall B, Hynes

AAAI Executive CouncilMeeting8:30 AM – 5:00 PM, Room 300, Hynes

◆Tutorial Program9:00 AM – 6:00 PM

Tutorial MA1: Applying Neural Network TechnologiesDavid Touretzky and Yann Le Cun9:00 AM – 1:00 PM, Ballroom A, Hynes

Tutorial MA2: Real-Time Knowledge-Based SystemsThomas Laffey, Cynthia Pickering, andN. S. Sridharan 9:00 AM – 1:00 PM, Ballroom C, Hynes

Tutorial MA3: Current Approach-es to Natural Language SemanticsGraeme Hirst and Chrysanne DiMarco9:00 AM – 1:00 PM, Room 302, Hynes

Tutorial MA4: Developing andManaging Expert Systems inBusiness and IndustryDavid Prerau and Earl Sacerdoti9:00 AM – 1:00 PM, Ballroom B, Hynes

Tutorial MA5: Truth Maintenance SystemsKen Forbus and Johan de Kleer9:00 AM – 1:00 PM, Room 304/6, Hynes

Tutorial MP1: Principles and Prac-tice of Knowledge AcquisitionThomas Gruber and Mark Musen2:00 PM – 6:00 PM, Ballroom C, Hynes

Tutorial MP2: Introduction toHypertext and HypermediaJakob Nielsen and Gerhard Fischer2:00 PM – 6:00 PM, Ballroom B, Hynes

Tutorial MP3: Evaluating Knowledge Engineering ToolsTod Loofbourrow and Randy Davis2:00 PM – 6:00 PM, Room 304/6, Hynes

Tutorial MP4: Building Blackboard ApplicationsDaniel Corkill and Rajendra Dodhiawala2:00 PM – 6:00 PM, Ballroom A, Hynes

Tutorial MP5: Case-Based ReasoningJanet Kolodner and Chris Riesbeck2:00 PM – 6:00 PM, Room 302, Hynes

Tuesday, July 31Conference Registration7:30 AM – 6:00 PM, Hall B, Hynes

AAAI Conference Committee Meeting7:30 AM, Room 300, Hynes

▲Plenary Addresses8:30 AM – 10:30 AM, Hynes Auditorium

IntroductionRaj Reddy, Past President, AAAI

AAAI Presidential Address: “Dimensions of Interaction”Daniel G. Bobrow, Xerox Palo AltoResearch Center

Why Is It So Hard Moving AI to Reality?Invited Speaker: Samuel Fuller, Vice Pres-ident, Research, Digital Equipment Cor-poration

Break10:30 AM – 11:00 AM

◆Invited Talk11:00 AM – 12:40 PM, Hynes Auditorium

The Future of Knowledge RepresentationSpeaker: Ronald J. Brachman, AT&T BellLaboratories

Knowledge Representation (KR) has tra-ditionally been thought of as the heartof AI. Anyone who has ever built anexpert system, a natural language sys-tem—almost any AI system at all—hashad to tackle the problem of represent-ing its knowledge of the world. Despiteits ubiquity, for most of AI’s history KRhas been a backstage activity. But inthe 1980s it emerged as a field untoitself, with its own burgeoning litera-ture and its own conference. Alongwith this growth, the last decade hasseen major changes in KR methodolo-gy, many important technical contribu-tions, and some dramatic challenges tothe basic assumptions of the field. Wewill survey some of these develop-ments, and then speculate about someof the equally interesting changes thatappear on the horizon. We will also

A–14 PROGRAM

Program

look at some of the critical problemsfacing KR research in the near future,both technical and sociological.

■ Technical Sessions11:00 AM – 12:40 PM

Automated Reasoning: Dis-tributed and Parallel SystemsSession Chair: Tom Dean11:00 AM – 12:40 PM, Ballroom A, Hynes

A Hierarchical Protocol for Coor-dinating Multiagent BehaviorsEdmund H. Durfee and Thomas A.Montgomery, University of Michigan11:00 AM – 11:25 AM, Ballroom A, Hynes

An Organizational Approach to Adaptive Production SystemsToru Ishida and Makoto Yokoo, NTTCommunications and Information Pro-cessing Laboratories; Les Gasser, University of Southern California11:25 AM – 11:50 AM, Ballroom A, Hynes

A Parallel Asynchronous Distributed Production SystemJames G. Schmolze and Suraj Goel,Tufts University11:50 AM – 12:15 PM, Ballroom A, Hynes

The Design of a Marker Passing Architecture for Knowledge ProcessingWing Lee and Dan Moldovan, University of Southern California12:15 PM – 12 40 PM, Ballroom A, Hynes

Machine Learning: Knowledge-Based InductiveLearning ISession Chair: Devika Subramanian11:00 AM – 12:40 PM, Ballroom B, Hynes

On Analytical and Similarity-Based ClassificationMarc Vilain, Phyllis Koton, and MelissaP. Chase, The MITRE Corporation11:00 AM – 11:25 AM, Ballroom B, Hynes

Learning from Textbook Knowledge: A Case Study William W. Cohen, Rutgers University11:25 AM – 11:50 AM, Ballroom B, Hynes

Changing the Rules: A Comprehensive Approach to Theory RefinementDirk Ourston and Raymond J. Mooney,University of Texas11:50 AM – 12:15 PM, Ballroom B, Hynes

Theory Reduction, Theory Revision, and RetranslationAllen Ginsberg, AT&T Bell Laboratories12:15 PM – 12:40 PM, Ballroom B, Hynes

Qualitative Reasoning: Reasoning with Multiple ModelsSession Chair: Brian Falkenhainer11:00 AM – 12:40 PM, Ballroom C, Hynes

Qualitative Reasoning withMicroscopic TheoriesShankar A. Rajamoney and Sang HoeKoo, University of Southern California11:00 AM – 11:25 AM, Ballroom C, Hynes

Shifting Ontological Perspectives in Reasoning about Physical SystemsZheng-Yang Liu and Arthur M. Farley,University of Oregon11:25 AM – 11:50 AM, Ballroom C, Hynes

Approximation ReformulationsDaniel Weld, University of Washington11:50 AM – 12:15 PM, Ballroom C, Hynes

Finding the Average Rates ofChange in Repetitive Behavior Alexander Yeh, Massachusetts Institute of Technology12:15 PM – 12:40 PM, Ballroom C, Hynes

Cognitive Modeling: Case-Based ReasoningSession Chair: Kevin Ashley11:00 AM – 12:40 PM, Room 304/6, Hynes

Distributed Cases for Case-BasedReasoning: Facilitating Use of Multiple CasesMichael Redmond, Georgia Institute ofTechnology11:00 AM – 11:25 AM, Room 304/6, Hynes

Integrating Planning and Acting in a Case-Based FrameworkKristian J. Hammond and TimothyConverse, University of Chicago11:25 AM – 11:50 AM, Room 304/6, Hynes

A Method of Calculating theMeasure of Salience in Understanding MetaphorsMakoto Iwayama, Takenobu Tokunaga,and Hozumi Tanaka, Tokyo Institute of Technology11:50 AM – 12:15 PM, Room 304/6, Hynes

Validated Retrieval in Case-Based ReasoningEvangelos Simoudis and James Miller,Brandeis University and Digital Equipment Corporation12:15 PM – 12:40 PM, Room 304/6, Hynes

■ Exhibits12:00 PM – 7:00 PM, Hall C & D, Hynes

Press Lunch and Conference12:00 PM, Room 208, Hynes

Lunch12:40 PM – 2:00 PM

Subgroup Meetings12:45 PM – 1:30 PM

AI in Business Subgroup Meeting12:45 PM – 1:30 PM, Room 302, Hynes

AI in Medicine Subgroup Meeting12:45 PM – 1:30 PM, Ballroom C, Hynes

SIGART Meeting12:45 PM – 1:30 PM, Room 304/6, Hynes

■ AI-On-Line2:00 PM – 4:00 PM, Room 210, Hynes

Six Companies Reveal Results of Expert Systems SuccessesOrganized by Chuck Williams, Infer-ence Corporation and Frank Angrisani,Sun America Financial. Moderator:Frank Angrisani, President, Sun Ameri-ca Financial. Panelists: Larry Tieman,American Airlines; Don McKay, Ameri-can President Companies, Roger Jam-bor, Dun & Bradstreet; John Woods,Ford Motor; Jim Euchner, Nynex; RichLevine, Sun America

SUNDAY – TUESDAY A–15

◆Invited Panel2:10 PM – 3:50 PM, Hynes Auditorium

AI & Software Engineering—Will the Twain Ever Meet?Moderator: Robert Balzer, University ofSouthern California–ISI. Panelists:Richard Fikes, Price Waterhouse; MarkFox, Carnegie Mellon University; JohnMcDermott, Digital Equipment Corpo-ration; Elliot Soloway, University ofMichigan

■ Technical Sessions2:10 PM – 3:50 PM

Natural Language: DiscourseSession Chair: Cecile Paris2:10 PM – 3:50 PM, Ballroom A, Hynes

PRAGMA—A Flexible Bidirectional Dialogue SystemJohn M. Levine, Univ. of Cambridge2:10 PM – 2:35 PM, Ballroom A, Hynes

Accent and Discourse Context: Assigning Pitch Accent in Synthetic SpeechJulia Hirschberg, AT&T Bell Laboratories2:35 PM – 3:00 PM, Ballroom A, Hynes

Structure of Perspectivity: A Case of Japanese Reflexive Pronoun “zibun”Yasuhiro Katagiri, NTT Basic Research Laboratories3:00 PM – 3:25 PM, Ballroom A, Hynes

Logical Task Modeling for Man-Machine DialogueM. D. Sadek, Centre National d’Etudesdes Télécommunications3:25 PM – 3:50 PM, Ballroom A, Hynes

Machine Learning: Connectionist MethodsSession Chair: Doug Fisher2:10 PM – 3:50 PM, Ballroom B, Hynes

Refinement of ApproximateDomain Theories by Knowledge-Based Artificial Neural NetworksGeoffrey G. Towell, Jude W. Shavlik,and Michiel O. Noordewier, Universityof Wisconsin2:10 PM – 2:35 PM, Ballroom B, Hynes

Empirical Studies on the Speed ofConvergence of Neural NetworkTraining Using Genetic AlgorithmsHiroaki Kitano, Carnegie Mellon Univ.2:35 PM – 3:00 PM, Ballroom B, Hynes

A Hybrid Connectionist, Symbolic Learning SystemLawrence O. Hall and Steve G. Romaniuk, University of South Florida3:00 PM – 3:25 PM, Ballroom B, Hynes

Explaining Temporal-Differencesto Create Useful Concepts forEvaluating StatesRichard C. Yee, Sharad Saxena, Paul E.Utgoff, and Andrew G. Barto, University of Massachusetts3:25 PM – 3:50 PM, Ballroom B, Hynes

Knowledge Representation: Causality and IntrospectionSession Chair: Peter Ladkin2:10 PM – 3:50 PM, Ballroom C, Hynes

Causal Theories for Nonmonotonic ReasoningHector Geffner, IBM T. J. WatsonResearch Center2:10 PM – 2:35 PM, Ballroom C, Hynes

A Circumscriptive Theory for Causal and Evidential SupportEunok Paek, Stanford University2:35 PM – 3:00 PM, Ballroom C, Hynes

A Formal Theory of MultipleAgent Nonmonotonic ReasoningLeora Morgenstern, IBM T. J. WatsonResearch Center3:00 PM – 3:25 PM, Ballroom C, Hynes

Decidable Reasoning in First-Order Knowledge Bases with Perfect IntrospectionGerhard Lakemeyer, Univ. of Toronto3:25 PM – 3:50 PM, Ballroom C, Hynes

Automated Reasoning: Truth Maintenance SystemsSession Chair: Jon Doyle2:10 PM – 3:50 PM, Room 304/6, Hynes

Exploiting Locality in a TMSJohan de Kleer, Xerox Palo AltoResearch Center2:10 PM – 2:35 PM, Room 304/6, Hynes

Computing Stable Models by Using the ATMSKave Eshghi, Hewlett-Packard Laboratories2:35 PM – 3:00 PM, Room 304/6, Hynes

Computing the Extensions ofAutoepistemic and Default Logicswith a Truth Maintenance SystemUlrich Junker, GMD; Kurt Konolige, SRI International3:00 PM – 3:25 PM, Room 304/6, Hynes

Maintaining Consistency in a Stra-tified Production System ProgramLouiqa Raschid, University of Maryland3:25 PM – 3:50 PM, Room 304/6, Hynes

Break3:50 PM – 4:40 PM

◆Invited Talk4:40 PM – 6:20 PM, Hynes Auditorium

Probably Approximately Correct LearningSpeaker: David Haussler, University ofCalifornia at Santa Cruz

This talk surveys recent theoreticalresults on the efficiency of machinelearning algorithms. The main tooldescribed is the notion of ProbablyApproximately Correct (PAC) learning,introduced by Valiant in 1984. In thefirst part of the talk, PAC learning isdefined and some examples and simpletheorems are given. In the second part,recent extensions and modifications ofthis model are outlined. These includedistribution specific results, learningwith queries, noise models, learningmulti-valued functions (including realand vector-valued functions), andincremental learning.


Automated Reasoning: Dis-tributed Artificial IntelligenceSession Chair: Bruce D’Ambrosio4:40 PM – 6:20 PM, Ballroom A, Hynes

On Acting TogetherHector J. Levesque and José H. T.Nunes, University of Toronto; Philip R.Cohen, SRI International4:40 PM – 5:05 PM, Ballroom A, Hynes

A–16 PROGRAM

Distributed Truth MaintenanceDavid Murray Bridgeland and MichaelN. Huhns, Microelectronics and Computer Technology Corporation5:05 PM – 5:30 PM, Ballroom A, Hynes

DARES: A Distributed AutomatedREasoning SystemS. E. Conry, D. J. MacIntosh and R. A.Meyer, Clarkson University5:30 PM – 5:55 PM, Ballroom A, Hynes

Negotiation and Conflict Resolution in Non-Cooperative DomainsGilad Zlotkin and Jeffrey S. Rosenschein, Hebrew University5:55 PM – 6:20 PM, Ballroom A, Hynes

Knowledge Acquisition andExpert Systems DesignMethodologiesSession Chair: Ramesh Patil4:40 PM – 6:20 PM, Ballroom B, Hynes

Establishing the Coherence of an Explanation to Improve Refinement of an Incomplete Knowledge BaseYoung-Tack Park and David C. Wilkins,University of Illinois4:40 PM – 5:05 PM, Ballroom B, Hynes

An Experiment in Direct Knowledge AcquisitionPeter W. Mullarkey, Schlumberger Laboratory for Computer Science5:05 PM – 5:30 PM, Ballroom B, Hynes

A Design Based Approach to Con-structing Computational Solutionsto Diagnostic ProblemsD. Volovik, I. A. Zaulkernan, P. E. John-son, University of Minnesota; C. E.Matthews, IBM Corporation5:30 PM – 5:55 PM, Ballroom B, Hynes

Parametric Engineering DesignUsing Constraint-Based ReasoningNiall Murtagh and Masamichi Shimura,Tokyo Institute of Technology5:55 PM – 6:20 PM, Ballroom B, Hynes

Knowledge Representation: Tem-poral and Spatial ReasoningSession Chair: Reid Simmons4:40 PM – 6:20 PM, Ballroom C, Hynes

Reasoning about QualitativeTemporal InformationPeter van Beek, University of Waterloo4:40 PM – 5:05 PM, Ballroom C, Hynes

Weak Representations of Interval AlgebrasGérard Ligozat, Université Paris XI5:05 PM – 5:30 PM, Ballroom C, Hynes

A Qualitative Model for SpaceAmitabha Mukerjee and Gene Joe,Texas A&M University5:30 PM – 5:55 PM, Ballroom C, Hynes

Solving Geometric Constraint SystemsGlenn A. Kramer, University of Sussexand Schlumberger Laboratory for Computer Science5:55 PM – 6:20 PM, Ballroom C, Hynes

Automated Reasoning: SearchSession Chair: Richard Korf4:40 PM – 6:20 PM, Room 304/6, Hynes

Path-Focused Duplication: ASearch Procedure for General MatingsSunil Issar, Carnegie Mellon University4:40 PM – 5:05 PM, Room 304/6, Hynes

Consistent Linear Speedups to aFirst Solution in Parallel State-Space SearchVikram A. Saletore and L. V. Kalé, Univ. of Illinois at Urbana-Champaign5:05 PM – 5:30 PM, Room 304/6, Hynes

Iterative BroadeningMatthew L. Ginsberg and William D.Harvey, Stanford University5:30 PM – 5:55 PM, Room 304/6, Hynes

Search Lessons Learned from Crossword PuzzlesMatthew L. Ginsberg, Michael Frank,Michael P. Halpin, and Mark C. Torrance, Stanford University5:55 PM – 6:20 PM, Room 304/6, Hynes

Reception7:00 PM – 9:00 PM, Marriott Grand Ballroom

Wednesday, August 1Conference Registration7:30 AM – 6:00 PM, Hall B, Hynes

The Artificial IntelligenceJournal Editorial Board7:00 AM, Vineyard Room, Marriott Copley Place

◆Invited Talk8:30 AM – 10:10 AM, Hynes Auditorium

Rationality and Its Roles in ReasoningSpeaker: Jon Doyle, Massachusetts Institute of Technology

The economic theory of rationalitypromises to equal mathematical logic inits importance for the mechanization ofreasoning. We survey the growing liter-ature on how the basic notions of prob-ability, utility, and rational choice, cou-pled with practical limitations on infor-mation and resources, influence thedesign and analysis of reasoning andrepresentation systems.

■ Technical Sessions8:30 AM – 10:10 AM

Intelligent Multimedia InterfacesSession Chair: Bob Neches8:30 AM – 10:10 AM, Ballroom A, Hynes

Understanding Natural Language with DiagramsGordon S. Novak Jr., University ofTexas at Austin and William C. Bulko,IBM Corporation8:30 AM – 8:55 AM, Ballroom A, Hynes

Avoiding Unwanted Conversational Implicatures in Text and GraphicsJoseph Marks and Ehud Reiter, Harvard University8:55 AM – 9:20 AM, Ballroom A, Hynes

Pointing: A Way Toward Explanation DialogueJohanna D. Moore, University of Pittsburgh; William R. Swartout,USC–Information Science Institute9:20 AM – 9:45 AM, Ballroom A, Hynes

TUESDAY – WEDNESDAY A–17

Coordinating Text and Graphicsin Explanation GenerationSteven K. Feiner and Kathleen R. McKeown, Columbia University9:45 AM – 10:10 AM, Ballroom A, Hynes

Machine Learning: Learningand Problem SolvingSession Chair: Prasad Tadepalli8:30 AM – 10:10 AM, Ballroom B, Hynes

Learning Abstraction Hierarchiesfor Problem SolvingCraig A. Knoblock, Carnegie MellonUniversity8:30 AM – 8:55 AM, Ballroom B, Hynes

Learning Search Control forConstraint-Based SchedulingMegan Eskey and Monte Zweben,NASA Ames Research Center8:55 AM – 9:20 AM, Ballroom B, Hynes

Adaptive Search by Explanation-Based Learning of Heuristics CensorsNeeraj Bhatnagar, Siemens CorporateResearch; Jack Mostow, Rutgers Univ.9:20 AM – 9:45 AM, Ballroom B, Hynes

Empirical Comparisons of SomeDesign Replay AlgorithmsBrad Blumenthal, University of Texas at Austin9:45 AM – 10:10 AM, Ballroom B, Hynes

Automated Reasoning: Constraint Satisfaction Problems ISession Chair: Alan Mackworth8:30 AM – 10:10 AM, Ballroom C, Hynes

Tree Decomposition with Appli-cations to Constraint ProcessingItay Meiri and Judea Pearl, Universityof California at Los Angeles; RinaDechter, Technion–Israel Institute of Technology 8:30 AM – 8:55 AM, Ballroom C, Hynes

Complexity of K-Tree StructuredConstraint Satisfaction ProblemsEugene C. Freuder, University of New Hampshire8:55 AM – 9:20 AM, Ballroom C, Hynes

Some Applications of GraphBandwidth to Constraint Satisfaction ProblemsRamin Zabih, Stanford University9:20 AM – 9:45 AM, Ballroom C, Hynes

The Complexity of ConstraintSatisfaction in PrologBernard A. Nadel, Wayne State Univ.9:45 AM – 10:10 AM, Ballroom C, Hynes

Perception and Signal Understanding: VisionSession Chair: Rod Brooks8:30 AM – 9:45 AM, Room 304/6, Hynes

Computing Exact Aspect Graphs of Curved Objects: Parametric SurfacesJean Ponce, University of Illinois; David J. Kriegman, Yale University8:30 AM – 8:55 AM, Room 304/6, Hynes

Generalized Shape AutocorrelationAndrea Califano and Rakesh Mohan,IBM T. J. Watson Research Center8:55 AM – 9:20 AM, Room 304/6, Hynes

Constraints for the Early Detection of Discontinuity from MotionMichael J. Black and P. Anandan, Yale University9:20 AM – 9:45 AM, Room 304/6, Hynes

Automated Reasoning: Theorem Proving and Program Synthesis ISession Chair: Richard Waters8:30 AM – 9:45 AM, Room 302, Hynes

Mechanizing Inductive ReasoningEmmanuel Kounalis and Michaël Rusinowitch, CRIN8:30 AM – 8:55 AM, Room 302, Hynes

Inductive Synthesis of Equational ProgramsNachum Dershowitz, Univ. of Illinois; Eli Pinchover, Bar-Ilan University8:55 AM – 9:20 AM, Room 302, Hynes

Automatically Generating Universal Attachments ThroughCompilationKaren L. Myers, Stanford University9:20 AM – 9:45 AM, Room 302, Hynes

■ AI-On-Line9:00 AM – 11:00 AM, Room 210, Hynes

Why and How CorporateAmerica Is Integrating AI into Data ProcessingOrganized by Jack Rahaim, Digital Equip-ment Corporation. Moderator: Don Mick,Aetna Life & Causalty. Panelists: TedSmith, US West; Rosemary Baer, Chemi-cal Bank; Paul Deschamps, Aetna Life andCausalty; Nancy Faucheux, Texaco Inc.

Exhibits10:00 AM – 6:00 PM, Halls C & D, Hynes

Break10:10 AM – 11:00 AM

◆Invited Panel11:00 AM – 12:40 PM, Hynes Auditorium

User Modeling and User InterfacesModerator: Kathleen McKeown,Columbia University. Panelists: Susan T.Dumais and James D. Hollan, Bellcore;Karen Sparck Jones, University of Cam-bridge; Candace Sidner, Digital Equip-ment Corporation


Automated Reasoning: Evidential ReasoningSession Chair: Mike Wellman11:00 AM – 12:40 PM, Ballroom A, Hynes

Two Views of Belief: Belief as Generalized Probabilityand Belief as EvidenceJoseph Y. Halpern and Ronald Fagin,IBM Almaden Research Center11:00 AM – 11:25 AM, Ballroom A, Hynes

The Belief Calculus and Uncertain ReasoningYen-Teh Hsia, Université Libre de Bruxelles11:25 AM – 11:50 AM, Ballroom A, Hynes

Symbolic Probabilistic Inferencein Belief NetworksRoss D. Shachter and Brendan A. DelFavero, Stanford University; BruceD’Ambrosio, Oregon State University11:50 AM – 12:15 PM, Ballroom A, Hynes

A–18 PROGRAM

Probabilistic Semantics for Cost-Based AbductionEugene Charniak and Solomon E. Shi-mony, Brown University12:15 PM – 12:40 PM, Ballroom A, Hynes

Machine Learning: InductiveLearning II—Relational andProbabilistic ModelsSession Chair: Wray Buntine11:00 AM – 12:40 PM, Ballroom B, Hynes

Effective Generalization of Relational DescriptionsLarry Watanabe and Larry Rendell,Univ. of Illinois at Urbana-Champaign11:00 AM – 11:25 AM, Ballroom B, Hynes

Inductive Learning in Probabilistic DomainYoichiro Nakakuki, Yoshiyuki Kosekiand Midori Tanaka, NEC Corporation11:25 AM – 11:50 AM, Ballroom B, Hynes

Learning Causal Trees from Dependence InformationDan Geiger, Northrop Research andTechnology Corporation; Azaria Paz,Israel Institute of Technology; JudeaPearl, Univ. of California at Los Angeles11:50 AM – 12:15 PM, Ballroom B, Hynes

Constructor: A System for theInduction of Probabilistic ModelsRobert M. Fung and Stuart L. Crawford,Advanced Decision Systems12:15 PM – 12:40 PM, Ballroom B, Hynes

Automated Reasoning: Planning ISession Chair: Mark Drummond11:00 AM – 12:40 PM, Ballroom C, Hynes

Incremental, Approximate PlanningCharles Elkan, University of Toronto11:00 AM – 11:25 AM, Ballroom C, Hynes

Synthesis of Reactive Plans forMulti-Path EnvironmentsFroduald Kabanza, Université de Liège11:25 AM – 11:50 AM, Ballroom C, Hynes

A Theory of Plan ModificationSubbarao Kambhampati, Stanford Univ.11:50 AM – 12:15 PM, Ballroom C, Hynes

Introducing the Tileworld:Experimentally Evaluating Agent ArchitecturesMartha E. Pollack, SRI International;Marc Ringuette, Carnegie Mellon Univ.12:15 PM – 12:40 PM, Ballroom C, Hynes

Knowledge Representation:InheritanceSession Chair: Phil Klahr11:00 AM – 12:40 PM, Room 304/6, Hynes

Terminological Cycles in KL-ONE-based Knowledge Representation LanguagesFranz Baader, German Research Centerfor Artificial Intelligence11:00 AM – 11:25 AM, Room 304/6, Hynes

A Temporal Terminological Logic Albrecht Schmiedel, Technische Universität Berlin11:25 AM – 11:50 AM, Room 304/6, Hynes

Boolean Extensions of Inheritance NetworksJohn F. Horty, University of Maryland;Richmond H. Thomason, University of Pittsburgh11:50 AM – 12:15 PM, Room 304/6, Hynes

On the Complexity of MonotonicInheritance with RolesRamiro A. de T. Guerreiro and AndreaS. Hemerly, IBM Brazil; Yoav Shoham,Stanford University12:15 PM – 12:40 PM, Room 304/6, Hynes

Theorem Proving II: TermRewritingSession Chair: David Poole11:00 AM – 11:50 AM, Room 302, Hynes

Skolem Functions and Equalityin Automated DeductionWilliam McCune, Argonne National Laboratory11:00 AM – 11:25 AM, Room 302, Hynes

Solving Term InequalitiesGerald E. Peterson, McDonnell Douglas Corporation11:25 AM – 11:50 AM, Room 302, Hynes

Lunch12:40 PM – 2:00 PM

AAAI Annual Business Meeting12:45 PM – 1:15 PM, Hynes Auditorium

AAAI Press Editorial Meeting1:20 PM, Room 300, Hynes


What Users are Buying Today inAI and Why: Reports from Man-agementOrganized by Keith Scovell and TimHolcomb, Andersen Consulting. Moder-ator: Roger Shelm, CIGNA. Panelists:Mary Dunn, Mutual of New York; BillCorbin, Owens Corning; Dave Wise,Frito-Lay; Mike Rogalski, Sears Mer-chandise Group; Troy A. Heindel,NASA Johnson Space Center; Al Brown,Lockheed Missiles & Space Co.

◆Tutorial Program2:00 pM – 6:00 PM

Tutorial WP1: Explanation-BasedLearning: Problems and MethodsSmadar Kedar and Steven Minton2:00 PM – 6:00 PM, Salon G, Marriott Copley Place

Tutorial WP2: AI and Engineering DesignDuvvuru Sriram and Chris Tong2:00 PM – 6:00 PM, Salon A, Marriott Copley Place

Tutorial WP3: Designing Complex Systems Using CLOSRichard Gabriel and John L. White2:00 PM – 6:00 PM, Salon F, Marriott Copley Place

Tutorial WP4: Qualitative Rea-soning About Physical SystemsBenjamin Kuipers and Elisha Sacks2:00 PM – 6:00 PM, Salon K, Marriott Copley Place

Tutorial WP5: Genetic Algo-rithms and Classifier SystemsDavid E. Goldberg and John R. Koza2:00 PM – 6:00 PM, Salon E, Marriott Copley Place

WEDNESDAY A–19


Developing Real-Time On-Line ApplicationsOrganized by Didi Murnane, GensymCorporation. Moderator: Karlene A.Kosanovich, E. I. DuPont de Nemoursand Company. Panelists: Dorothy Yu,Coopers & Lybrand; Karl-Erik Arzen,Lund Institute of Technology, Sweden;Gregory M. O'Connor, MassachusettsInstitute of Technology; Paul D.Schoen, Rockwell International; Jean-Pierre Aubert, Alcatel ISR France.

Thursday, August 2Conference Registration7:30 AM – 6:00 PM, Hall B, Hynes

AAAI Publication Committee Meeting7:30 AM, Room 300, Hynes


Truth Maintenance Speaker: David A. McAllester, Mas-sachusetts Institute of Technology

A survey of the state of the art in truthmaintenance with an emphasis on cur-rent theoretical research issues.


Natural Language: InterpretationSession Chair: Julia Hirschberg8:30 AM – 10:10 AM, Ballroom A, Hynes

Towards Incremental Disam-biguation with a GeneralizedDiscrimination NetworkManabu Okumura and Hozumi Tana-ka, Tokyo Institute of Technology8:30 AM – 8:55 AM, Ballroom A, Hynes

Parsing a Natural Language UsingMutual Information StatisticsDavid M. Magerman and Mitchell P.Marcus, University of Pennsylvania8:55 AM – 9:20 AM, Ballroom A, Hynes

Integrating Natural Language Processing and Knowledge Based ProcessingRebecca Passonneau, Carl Weir, andTim Finin, Unisys Corporation; MarthaPalmer, Center for Advanced Information Technology9:20 AM – 9:45 AM, Ballroom A, Hynes

Truly Parallel Understanding of TextYeong-Ho Yu and Robert F. Simmons,University of Texas at Austin9:45 AM – 10:10 AM, Ballroom A, Hynes

Machine Learning: Inductive Learning IIISession Chair: Philip Laird8:30 AM – 10:10 AM, Ballroom B, Hynes

Adding Domain Knowledge toSBL through Feature ConstructionChristopher John Matheus, GTE Laboratories Incorporated8:30 AM – 8:55 AM, Ballroom B, Hynes

What Should Be Minimized in a Decision Tree?Usama M. Fayyad and Keki B. Irani,University of Michigan8:55 AM – 9:20 AM, Ballroom B, Hynes

Myths and Legends in Learning Classification RulesWray Buntine, Turing Institute9:20 AM – 9:45 AM, Ballroom B, Hynes

Inductive Learning in a Mixed Paradigm SettingDavid B. Skalak and Edwina L. Rissland,University of Massachusetts9:45 AM – 10:10 AM, Ballroom B, Hynes

Robotics: Robot PlanningSession Chair: Pattie Maes8:30 AM – 10:10 AM, Ballroom C, Hynes

Symmetry Constraint Inference inAssembly Planning: Automatic As-sembly Configuration SpecificationYanxi Liu and Robin J. Popplestone,University of Massachusetts at Amherst8:30 AM – 8:55 AM, Ballroom C, Hynes

Indexical Knowledge in Robot PlansYves Lespérance and Hector J.Levesque, University of Toronto8:55 AM – 9:20 AM, Ballroom C, Hynes

A Hierachical Planner that Generates its Own HierachiesJens Christensen, Stanford University9:20 AM – 9:45 AM, Ballroom C, Hynes

Learning General Completable Reactive PlansMelinda T. Gervasio, University of Illinois at Urbana-Champaign9:45 AM – 10:10 AM, Ballroom C, Hynes

Knowledge Representation: Representation and UncertaintySession Chair: Kurt Konolige8:30 AM – 10:10 AM, Room 304/6, Hynes

A Maximum Entropy Approachto Nonmonotonic ReasoningMoisés Goldszmidt and Judea Pearl,University of California at Los Angeles;Paul Morris, IntelliCorp8:30 AM – 8:55 AM, Room 304/6, Hynes

A Probabilistic Interpretation forLazy Nonmonotonic ReasoningKen Satoh, Institute for New Generation Computer Technology8:55 AM – 9:20 AM, Room 304/6, Hynes

A Hybrid Framework for Repre-senting Uncertain KnowledgeAlessandro Saffiotti, Université Libre de Bruxelles9:20 AM – 9:45 AM, Room 304/6, Hynes

Probabilities that Imply CertaintiesHaim Shvaytser, David Sarnoff Research Center9:45 AM – 10:10 AM, Room 304/6, Hynes

■ AI-On-Line9:00 AM – 11:00 AM, Room 210, Hynes

Differentiating Expert Systems ProductsOrganized by Larry Harris, AICorp andRobert Keller, Renaissance Internation-al. Moderator: E. Robert Keller, Presi-dent, Renaissance International. Pan-elists: Larry Harris, AICorp; Harry C.Reinstein, Aion Corp.; Alain Rappaport,Neuron Data; Robert Moore, Gensym.

A–20 PROGRAM

■ Exhibits10:00 AM – 5:00 PM, Hall C & D, Hynes

Break10:10 AM – 11:00 AM

▼Highlights from the Inno-vative Applications of Artificial Intelligence Con-ference, May 1990—Part IModerator: Howard Shrobe, StandingConference Committee Chair11:00 AM – 12:15 PM, Hynes Auditorium

Inspector: An Expert System ForMonitoring Worldwide TradingActivities in Foreign ExchangeElizabeth Byrnes, Thomas Campfield,Neil Henry, and Steven Waldman,Manufacturers Hanover Trust11:00 AM – 11:25 AM, Hynes Auditorium

CONSTRUE/TIS: A System forContent-Based Indexing of aDatabase of News StoriesPhilip J. Hayes, Carnegie Group Incorpo-rated; Steven P. Weinstein, Reuters Ltd11:25 AM – 11:50 AM, Hynes Auditorium

A Real-Time Alarm Analysis AdvisorSteven Silverman, James Dixon, TimFink, and Paul Kotas, Consolidated Edi-son Company of New York, Incorporat-ed; Alvin Shoop, Bhashyam Ramesh,and Philip Klahr, Inference Corporation11:50 AM – 12:15 PM, Hynes Auditorium


Intelligent Interfaces and Plan RecognitionSession Chair: Tom Malone11:00 AM – 12:40 PM, Ballroom A, Hynes

Models of Plans to Support Com-munication: An Initial ReportKaren E. Lochbaum and Barbara J.Grosz, Harvard University; Candace L.Sidner, Digital Equipment Corporation11:00 AM – 11:25 AM, Ballroom A, Hynes

Incorporating Default Inferences Into Plan RecognitionSandra Carberry, University of Delaware11:25 AM – 11:50 AM, Ballroom A, Hynes

A Cooperative Problem SolvingSystem for User Interface Design Andreas C. Lemke and Gerhard Fischer,University of Colorado11:50 AM – 12:15 PM, Ballroom A, Hynes

A Collaborative Interface forEditing Large Knowledge BasesLoren G. Terveen and David A. Wroblew-ski, MCC Human Interface Laboratory12:15 PM – 12:40 PM, Ballroom A, Hynes

Commonsense Reasoning: Model-Based Diagnosis and Design ISession Chair: Benjamin Kuipers11:00 AM – 12:15 PM, Ballroom B, Hynes

Interaction-Based Invention:Designing Novel Devices from First PrinciplesBrian Williams, Xerox Palo Alto Research Center11:00 AM – 11:25 AM, Ballroom B, Hynes

Characterizing DiagnosesJohan de Kleer, Xerox Palo AltoResearch Center; Alan K. Mackworth,University of British Columbia; Ray-mond Reiter, University of Toronto11:25 AM – 11:50 AM, Ballroom B, Hynes

Abductive and Default Reason-ing: A Computational CoreBart Selman and Hector J. Levesque,University of Toronto11:50 AM – 12:15 PM, Ballroom B, Hynes

Automated Reasoning: Planning IISession Chair: Daniel Corkill11:00 AM – 12:40 PM, Ballroom C, Hynes

ABTWEAK: Abstracting a Nonlin-ear, Least Commitment PlannerQiang Yang, University of Waterloo;Josh D. Tenenberg, Univ. of Rochester11:00 AM – 11:25 AM, Ballroom C, Hynes

The STRIPS Assumption for Planning Under UncertaintyMichael P. Wellman, Wright-Patterson AFB11:25 AM – 11:50 AM, Ballroom C, Hynes

Getting Serious About ParsingPlans: A Grammatical Analysis ofPlan RecognitionMarc Vilain, The MITRE Corporation11:50 AM – 12:15 PM, Ballroom C, Hynes

Mapping and Retrieval DuringPlan Reuse: A Validation Structure Based ApproachSubbarao Kambhampati, Stanford Univ.12:15 PM – 12:40 PM, Ballroom C, Hynes

Knowledge Representation:Complexity and ExpressivenessSession Chair: Henry Kautz11:00 AM – 12:40 PM, Room 304/6, Hynes

An Optimally Efficient Limited Inference SystemLokendra Shastri and Venkat Ajjana-gadde, University of Pennsylvania11:00 AM – 11:25 AM, Room 304/6, Hynes

On the Expressiveness of Networks with Hidden VariablesRina Dechter, Technion–Israel Institute of Technology11:25 AM – 11:50 AM, Room 304/6, Hynes

The Complexity of Closed WorldReasoning and CircumscriptionMarco Cadoli and Maurizio Lenzerini,Università di Roma “La Sapienza”11:50 AM – 12:15 PM, Room 304/6, Hynes

It’s Not My Default: The Complexity of MembershipProblems in Restricted Propositional Default LogicsJonathan Stillman, General ElectricResearch and Development Center12:15 PM – 12:40 PM, Room 304/6, Hynes

Lunch12:40 PM – 2:00 PM

Subgroup Meetings12:45 PM – 1:30 PM

AI and the Law Subgroup Meeting12:45 PM – 1:30 PM, Room 302, Hynes

AI in Manufacturing Subgroup Meeting12:45 PM – 1:30 PM, Room 304/6, Hynes

THURSDAY A–21

■ AI-on-Line2:00 PM – 4:00 PM, Room 210, Hynes

AI Versus Conventional DataProcessingOrganized by Keith Scovell and TimHolcomb, Andersen Consulting. Mod-erator: Bruce B. Johnson, AndersonConsulting. Panelists: Bradley Billet-deaux, Exxon Company USA; StanleyWozniak, MCI; Dave Dean, EastmanKodak; Glen Galen, Burlington North-ern Railroad Information Systems, EdMahler, E. I. DuPont de Nemours andCompany

▼Highlights from the Inno-vative Applications of Arti-ficial Intelligence Confer-ence, May 1990—Part IIModerator: Howard Shrobe, StandingConference Committee Chair2:10 AM – 3:25 PM, Hynes Auditorium

Prism: A Case-Based Telex ClassifierMarc Goodman, Cognitive Systems Incorporated2:10 PM – 2:35 PM, Hynes Auditorium

National Dispatcher Router:AMulti-Paradigm Based Schedul-ing Advisor or From Prototype toProduction SystemJanet Rothstein, Digital Equipment Corporation2:35 PM – 3:00 PM, Hynes Auditorium

Development of Expert SystemsSupported Construction Planningfor Shield Tunneling MethodMinoru Harada and Zenichi Igarashi,Okumura Corporation; Satoshi Okuide,Hitachi Ltd; Yasuhiro Kitagawa, HitachiSeibu Software Company Ltd3:00 PM – 3:25 PM, Hynes Auditorium


Commonsense Reasoning: Qualitative Modeling of Physical SystemsSession Chair: Leo Joskowicz2:10 PM – 3:50 PM, Ballroom A, Hynes

QPC: A Compiler from PhysicalModels into Qualitative Differential EquationsJ. Crawford, A. Farquhar and B. Kuipers,University of Texas at Austin2:10 PM – 2:35 PM, Ballroom A, Hynes

Self-Explanatory Simulations: An Integration of Qualitative and Quantitative KnowledgeKenneth D. Forbus, University of Illinois; Brian Falkenhainer, Xerox Palo Alto Research Center2:35 PM – 3:00 PM, Ballroom A, Hynes

Dynamic Across-Time Measurement InterpretationDennis DeCoste, University of Illinois3:00 PM – 3:25 PM, Ballroom A, Hynes

Obtaining Quantitative Predictions from Monotone RelationshipsJoseph Hellerstein, IBM T. J. WatsonResearch Center3:25 PM – 3:50 PM, Ballroom A, Hynes

Machine Learning: Discovery and Learning RobotsSession Chair: Tom Mitchell2:10 PM – 3:50 PM, Ballroom C, Hynes

Learning to Coordinate BehaviorsPattie Maes and Rodney A. Brooks,Massachusetts Institute of Technology2:10 PM – 2:35 PM, Ballroom C, Hynes

Two Case Studies in Cost-Sensitive Concept AcquisitionMing Tan and Jeffrey C. Schlimmer,Carnegie Mellon University2:35 PM – 3:00 PM, Ballroom C, Hynes

A Proven Domain-Independent ScientificFunction-Finding AlgorithmCullen Schaffer, Rutgers University3:00 PM – 3:25 PM, Ballroom C, Hynes

Automated Discovery in a Chemistry LaboratoryJan M. Zytkow, Jieming Zhu and AbulHussam, George Mason University3:25 PM – 3:50 PM, Ballroom C, Hynes

Automated Reasoning: Constraint Satisfaction Problems IISession Chair: David McAllester2:10 PM – 3:25 PM, Room 302, Hynes

Dynamic Constraint Satisfaction ProblemsSanjay Mittal and Brian Falkenhainer,Xerox Palo Alto Research Center2:10 PM – 2:35 PM, Room 302, Hynes

An Algebraic Approach to Conflict Resolution in PlanningQiang Yang, University of Waterloo2:35 PM – 3:00 PM, Room 302, Hynes

Solving Large-Scale Constraint-Satisfaction andScheduling Problems Using a Heuristic Repair MethodSteven Minton, Andrew B. Philips andPhilip Laird, NASA Ames Research Center; Mark D. Johnston, Space Telescope Science Institute3:00 PM – 3:25 PM, Room 302, Hynes

Knowledge Representation: Default RepresentationsSession Chair: Bill Mark2:10 PM – 3:50 PM, Room 304/6, Hynes

The Representation of Defaults in CycRamanathan V. Guha, MCC2:10 PM – 2:35 PM, Room 304/6, Hynes

Conditional Logics of Normality as Modal SystemsCraig Boutilier, University of Toronto2:35 PM – 3:00 PM, Room 304/6, Hynes

The Generalized Theory of Model PreferencePiotr Rychlik, Polish Acad. of Sciences3:00 PM – 3:25 PM, Room 304/6, Hynes

Nonmonotonicity and the Scopeof Reasoning: Preliminary ReportDavid W. Etherington, AT&T Bell Labo-ratories; Sarit Kraus and Donald Perlis,University of Maryland3:25 PM – 3:50 PM, Room 304/6, Hynes

Break3:50 PM – 4:40 PM

A–22 PROGRAM

◆Invited Talk4:40 PM – 6:20 PM, Hynes Auditorium

Reasoning and Acting in Real TimeSpeaker: Stanley J. Rosenschein, Teleos Research

Some reasoning problems can beapproached as if time didn’t matter:facts are presented to the system, con-clusions are derived, and the usefulnessof the conclusions is relatively inde-pendent of the time it takes to derivethem. In many real-world reasoningproblems, however, time is an essentialconsideration: the world is changing,new information is being acquired, andreasoning processes must produceactions that can be performed whilethey are still timely. Unfortunately,many of the symbolic reasoning tech-niques developed by the artificial intel-ligence research community over theyears are more appropriate to “time-less” reasoning than to “timely” rea-soning. Recently, however, a new bodyof AI research has sprung up thatdirectly addresses the issue of reason-ing and acting under time pressure.This talk will survey this new researcharea and suggest directions for futurework.


AI and EducationSession Chair: Beverly Woolf4:40 PM – 5:55 PM, Ballroom A, Hynes

Backward Model Tracing: AnExplanation-Based Approach forReconstructing Student ReasoningDanilo Fum, Università di Trieste; Paolo Giangrandi, and Carlo Tasso, Università di Udine4:40 PM – 5:05 PM, Ballroom A, Hynes

A Blackboard-Based DynamicInstructional PlannerWilliam R. Murray, FMC CorporateTechnology Center5:05 PM – 5:30 PM, Ballroom A, Hynes

Towards a System Architecture Supporting Contextualized LearningGerhard Fischer, Andreas C. Lemke, andRaymond McCall, Univ. of Colorado5:30 PM – 5:55 PM, Ballroom A, Hynes

Machine Learning: Speed Up LearningSession Chair: Tom Ellman4:40 PM – 6:20 PM, Ballroom B, Hynes

Extending EBG to Term-Rewriting SystemsPhilip Laird and Evan Gamble, NASAAmes Research Center4:40 PM – 5:05 PM, Ballroom B, Hynes

Why Prodigy/EBL WorksOren Etzioni, Carnegie Mellon Univ.5:05 PM – 5:30 PM, Ballroom B, Hynes

Operationality Criteria for Recursive PredicatesStanley Letovsky, Carnegie Mellon Univ.5:30 PM – 5:55 PM, Ballroom B, Hynes

The Utility of EBL in Recursive Domain TheoriesDevika Subramanian and Ronen Feld-man, Cornell University5:55 PM – 6:20 PM, Ballroom B, Hynes

Automated Reasoning: Planning IIISession Chair: David Wilkins4:40 PM – 6:20 PM, Ballroom C, Hynes

An Approach to ReasoningAbout Continuous Change forApplications in PlanningThomas Dean and Greg Siegle, Brown University4:40 PM – 5:05 PM, Ballroom C, Hynes

Anytime Synthetic Projection:Maximizing the Probability ofGoal SatisfactionMark Drummond and John Bresina,NASA Ames Research Center5:05 PM – 5:30 PM, Ballroom C, Hynes

Admissible Criteria for LoopControl in PlanningRoy Feldman and Paul Morris, IntelliCorp5:30 PM – 5:55 PM, Ballroom C, Hynes

Practical Temporal ProjectionSteven Hanks, Univ. of Washington5:55 PM – 6:20 PM, Ballroom C, Hynes

Knowledge Representation: Databases and ConnectionismSession Chair: James Schmolze4:40 PM – 5:55 PM, Room 304/6, Hynes

The Intelligent Database Interface: Integrating AI and Database SystemsDon McKay, Tim Finin, and AnthonyO’Hare, Unisys Center for AdvancedInformation Technology4:40 PM – 5:05 PM, Room 304/6, Hynes

A Structured Connectionist Unification AlgorithmSteffen Hölldobler, International Computer Science Institute5:05 PM – 5:30 PM, Room 304/6, Hynes

Connectionism, Rule Following,and Symbolic ManipulationRobert F. Hadley, Simon Fraser Univ.5:30 PM – 5:55 PM, Room 304/6, Hynes

Tour of MIT AI Lab 5:00 PM – 7:30 PM

Friday, August 3Conference Registration7:30 AM – 11:30 AM, Hall B, Hynes

IJCAI Trustee’s Meeting8:00 AM – 6:00 PM, Vineyard Room,Marriott Copley Place


Massively Parallel AISpeaker: David L. Waltz, ThinkingMachines Corporation and BrandeisUniversity

It is argued that massively parallelmachines (e. g. the ConnectionMachine) offer a better match for cog-nitive as well as for applied AI thanserial computers. However, existing AIparadigms need to be modified orreplaced to realize the benefits, and

THURSDAY – FRIDAY A–23

some traditional AI problems (e. g.expert systems with small numbers ofrules) may not be large enough to justi-fy a massively parallel solution. A briefhistory, a description of some newparadigms and practical examples, anda discussion of future prospects formassively parallel AI will be included.


Commonsense Reasoning: Model-Based Diagnosis IISession Chair: Walter Hamscher8:30 AM – 10:10 AM, Ballroom A, Hynes

On the Role of Coherence inAbductive ExplanationHwee Tou Ng and Raymond J. Mooney,University of Texas at Austin8:30 AM – 8:55 AM, Ballroom A, Hynes

Model-Based Diagnosis of Planning FailuresLawrence Birnbaum, Gregg Collins,Michael Freed, and Bruce Krulwich,Northwestern University8:55 AM – 9:20 AM, Ballroom A, Hynes

Efficient Diagnosis of Multiple Disorders Based on aSymptom Clustering ApproachThomas D. Wu, Massachusetts Institute of Technology9:20 AM – 9:45 AM, Ballroom A, Hynes

Physical Impossiblity Instead of Fault ModelsGerhard Friedrich, Georg Gottlob,Wolfgang Nejdl, Technical University of Vienna9:45 AM – 10:10 AM, Ballroom A, Hynes

Machine Learning: General-ization and SpecializationSession Chair: Paul Utgoff8:30 AM – 10:10 AM, Ballroom B, Hynes

Generalization with Taxonomic InformationAlan M. Frisch and C. David Page, University of Illinois8:30 AM – 8:55 AM, Ballroom B, Hynes

Complementary DiscriminationLearning: A Duality Between Gen-eralization and DiscriminationWei-Min Shen, MCC8:55 AM – 9:20 AM, Ballroom B, Hynes

Incremental Non-BacktrackingFocusing: A Polynomially Bound-ed Generalization Algorithm for Version SpacesBenjamin D. Smith and Paul S. Rosen-bloom, Univ. of Southern California9:20 AM – 9:45 AM, Ballroom B, Hynes

Knowledge Level and InductiveUses of Chunking (EBL)Paul S. Rosenbloom, University ofSouthern California - ISI; Jans Aasman,Rijksuniversiteit Groningen9:45 AM – 10:10 AM, Ballroom B, Hynes

Robotics: Intelligent Mobile RobotsSession Chair: Stan Rosenschein8:30 AM – 10:10 AM, Ballroom C, Hynes

Integrating Execution, Planning,and Learning in Soar for External EnvironmentsJohn E. Laird, University of Michigan;Paul S. Rosenbloom, University ofSouthern California–ISI8:30 AM – 8:55 AM, Ballroom C, Hynes

Becoming Increasingly ReactiveTom M. Mitchell, Carnegie Mellon Univ.8:55 AM – 9:20 AM, Ballroom C, Hynes

Coping with Uncertainty in a Control System for Navigation and ExplorationThomas Dean, Kenneth Basye, RobertChekaluk, Seungseok Hyun, MoisesLejter and Margaret Randazza, BrownUniversity9:20 AM – 9:45 AM, Ballroom C, Hynes

LOGnets: A Hybrid Graph Spatial Representation for Robot NavigationPeter K. Malkin and Sanjaya Addanki,IBM T. J. Watson Research Center9:45 AM – 10:10 AM, Ballroom C, Hynes

Knowledge Representation:Architectures Session Chair: Bob MacGregor8:30 AM – 10:10 AM, Room 304/6, Hynes

On the Performance of LazyMatching in Production SystemsDaniel P. Miranker, David Brant, BernieLofaso and David Gadbois, University of Texas at Austin8:30 AM – 8:55 AM, Room 304/6, Hynes

A Framework for Investigating Production System Formulations with Polynomially Bounded MatchMilind Tambe, Carnegie Mellon Uni-versity; Paul S. Rosenbloom, Universityof Southern California - ISI8:55 AM – 9:20 AM, Room 304/6, Hynes

Very Fast Decision Table Execution of Propositional Expert SystemsRobert M. Colomb and Charles Y. C.Chung, CSIRO Division of InformationTechnology9:20 AM – 9:45 AM, Room 304/6, Hynes

A Principled Approach to Reasoning About the Specificity of RulesJohn Yen, Texas A&M University9:45 AM – 10:10 AM, Room 304/6, Hynes

Break10:10 AM – 11:00 AM

▲Plenary Addresses11:00 AM – 12:40 PM, Hynes Auditorium

IntroductionDaniel G. Bobrow, President AAAI

AI and the National Technologi-cal InfrastructureCraig Fields, Department of Defense

A–24 PROGRAM

A Blackboard-Based Dynamic InstructionalPlanner, A-23

A Real-Time Alarm Analysis Advisor, A-21A-24, John E., A-23Aasman, Jans, A-24Abductive and Default Reasoning: A Compu-

tational Core, A-21ABTWEAK: Abstracting a Nonlinear Least

Commitment Planner, A-21Accent and Discourse Context: Assigning

Pitch Accent in Synthetic Speech, A-16Adaptive Search by Explanation-Based Learn-

ing of Heuristic Censors, A-18Addanki, Sanjaya, A-24Adding Domain Knowledge to SBL Through

Feature Construction, A-20Addison, Edwin, A-14Admissible Criteria for Loop Control in Plan-

ning, A-23AI & Software Engineering—Will the Twain

Ever Meet? A-16AI and Engineering Design, A-19AI and Software Engineering: Will the Twain

Ever Meet?, A-16AI and Software Engineering— Managing

Exploratory Programming, A-16AI and the National Technological Infrastruc-

ture, A-24AI Versus Conventional Data Processing, A-7Ajjanagadde, Venkat, A-21Algebraic Approach to Conflict Resolution in

Planning, An, A-22Anandan, P., A-18Angrisani, Frank, A-2, A-15Anytime Synthetic Projection: Maximizing

the Probability of Goal Satisfaction, A-23Applying Neural Network Technologies, A-14Approach to Reasoning About Continuous

Change for Applications in Planning, An,A-23

Approximation Reformulations, A-15Arzen, Karl-Erik, A-5, A-20Aubert, Jean-Pierre, A-3, A-20Aubert, Jean-Pierre, A-6, A-20Automated Discovery in a Chemistry Labora-

tory, A-22Automatically Generating Universal Attach-

ments Through Compilation, A-18Avoiding Unwanted Conversational Implica-

tures in Text and Graphics, A-17Baader, Franz, A-19Backward Model Tracing: An Explanation-

Based Approach for Reconstructing StudentReasoning, A-23

Baer, Rosemary, A-3, A-18Balzer, Robert, A-16Barto, Andrew G., A-16Basic Theory of Neural Networks, A-14Basye, Kenneth, A-24Becoming Increasingly Reactive, A-24Belief Calculus and Uncertain Reasoning,

The, A-18Bhatnagar, Neeraj, A-18Billetdeaux, Bradley C., A-7, A-22Birnbaum, Lawrence, A-24Black, Michael J., A-18Blumenthal, Brad, A-18

Bobrow, Daniel G., A-14Boolean Extensions of Inheritance Networks,

A-19Boutilier, Craig, A-22Brachman, Ronald J., A-14Brant, David, A-24Bresina, John, A-23Bridgeland, David Murray, A-17Brooks, Rodney A., A-22Brown, Al, A-5, A-19Building Blackboard Applications, A-14Bulko, William C., A-17Buntine, Wray, A-20Byrnes, Elizabeth, A-21Cadoli, Marco, A-21Califano, Andrea, A-18Campfield, Thomas, A-21Carberry, Sandra, A-21Case-Based Reasoning, A-14Causal Theories for Nonmonotonic Reason-

ing, A-16Changing the Rules: A Comprehensive

Approach to Theory Refinement, A-15Characterizing Diagnoses, A-21Charniak, Eugene, A-19Chase, Melissa P., A-15Chekaluk, Robert, A-24Christensen, Jens, A-20Chung, Charles Y. C., A-24Circumscriptive Theory for Causal and Evi-

dential Support, A, A-16Cohen, Philip R., A-16Cohen, William W., A-15Collaborative Interface for Editing Large

Knowledge Bases, A, A-21Collins, Gregg, A-24Colomb, Robert M., A-24Complementary Discrimination Learning: A

Duality Between Generalization and Dis-crimination, A-24

Complexity of Closed World Reasoning andCircumscription, The, A-21

Complexity of Constraint Satisfaction in Pro-log, The, A-18

Complexity of K-Tree Structured Constraint Sat-isfaction Problems, A-18

Computing Exact Aspect Graphs of CurvedObjects: Parametric Surfaces, A-18

Computing Stable Models by Using theATMS, A-16

Computing the Extensions of Autoepistemicand Default Logics with a Truth Mainte-nance System, A-16

Conditional Logics of Normality as ModalSystems, A-22

Connectionism, Rule Following, and SymbolicManipulation, A-23

Conry, S.E., A-17Consistent Linear Speedups to a First Solu-

tion in Parallel State-Space Search, A-17Constraint Reasoning: Theory and Applica-

tions, A-14Constraints for the Early Detection of Dis-

continuity from Motion, A-18Constructor: A System for the Induction of

Probabilistic Models, A-19

CONSTRUE/TIS: A System for Content-BasedIndexing of a Database of News Stories, A-21

Converse, Timothy, A-15Cooperative Problem Solving System for User

Interface Design, A, A-21Coordinating Text and Graphics in Explana-

tion Generation, A-18Coping with Uncertainty in a Control System

for Navigation and Exploration, A-24Corbin, Bill, A-4, A-19Corkill, Daniel, A-14Crawford, J., A-22Crawford, Stuart L., A-19Current Approaches to Natural Language

Semantics, A-14D’Ambrosio, Bruce, A-18DARES: A Distributed Automated REasoning

System, A-17Davis, Randy, A-14de Kleer, Johan, A-14, A-16, A-21Dean, Dave, A-7, A-22Dean, Thomas, A-23, A-24Dechter, Rina, A-18, A-21Decidable Reasoning in First-Order Knowl-

edge Bases with Perfect Introspection, A-16DeCoste, Dennis, A-22Del Favero, Brendan A., A-18Dershowitz, Nachum, A-18Deschamps, Paul, A-3, A-18Design Based Approach to Constructing

Computational Solutions to DiagnosticProblems, A, A-17

Design of a Marker Passing Architecture forKnowledge Processing, The, A-15

Designing Complex Systems Using CLOS, A-19

Developing and Managing Expert Systems inBusiness and Industry, A-14

Developing Real-Time On-Line Applications,A-5

Developing Software Is Like Talking to Eskimo’sAbout Snow, A-16

Development of Expert Systems SupportedConstruction Planning for Shield Tunnel-ing Method, A-22

Differentiating Expert Systems Products, A-6DiMarco, Chrysanne, A-14Dimensions of Interaction, A-14Distributed Artificial Intelligence, A-14Distributed Cases for Case-Based Reasoning;

Facilitating Use of Multiple Cases, A-15Distributed Truth Maintenance, A-17Dixon, James, A-21Dodhiawala, Rajendra, A-14Doyle, Jon, A-17Drummond, Mark, A-23Dumais, Susan T., A-18Dunn, Mary, A-4, A-19Durfee, Edmund H.,A-15Dynamic Across-Time Measurement Interpre-

tation, A-22Dynamic Constraint Satisfaction Problems,

A-22Effective Generalization of Relational

Descriptions, A-19

AAAI ‘90 A–25

Index

Efficient Diagnosis of Multiple Disorders Basedon a Symptom Clustering Approach, A-24

Elkan, Charles, A-19Empirical Comparisons of Some Design

Replay Algorithms, A-18Empirical Studies on the Speed of Conver-

gence of Neural Network Training UsingGenetic Algorithms, A-16

Eshghi, Kave, A-16Eskey, Megan, A-18Establishing the Coherence of an Explana-

tion to Improve Refinement of an Incom-plete Knowledge Base, A-17

Etherington, David W., A-22Etzioni, Oren, A-23Euchner, Jim A-2, A-15Evaluating Knowledge Engineering Tools, A-

14Experiment in Direct Knowledge Acquisition,

An, A-17Explaining Temporal Differences to Create

Useful Concepts for Evaluating States, A-16

Explanation-Based Learning: Problems andMethods, A-19

Exploiting Locality in a TMS, A-16Extending EBG to Term-Rewriting Systems,

A-23Fagin, Ronald, A-18Falkenhainer, Brian, A-22Farley, Arthur M., A15Farquhar, A., A-22Faucheux, Nancy, A-3, A-18Fayyad, Usama M., A-20Feiner, Steven K., A-18Feldman, Ronen, A-23Feldman, Roy, A-23Fields, Craig, A-24Fikes, Richard, A-16Finding the Average Rates of Change in Repeti-

tive Behavior, A-15Finin, Tim, A-23, A-20Fink, Tim, A-21Fischer, Gerhard, A-14, A-21, A-23Forbus, Ken, A-14Forbus, Kenneth D., A-22Formal Theory of Multiple Agent Nonmono-

tonic Reasoning, A, A-16Fox, Mark S., A-16Framework for Investigating Production Sys-

tem Formulations with PolynomiallyBounded Match, A, A-24

Frank, Michael, A-17Freed, Michael, A-24Freuder, Eugene C., A-18Friedrich, Gerhard, A-24Frisch, Alan M., A-24Fuller, Samuel, A-14Fum, Danilo, A-23Fung, Robert M., A-19Future of Knowledge Representation, The, A-

14Gabriel, Richard, A-19Gadbois, David, A-24Galen, Glen, A-7, A-22Gamble, Evan, A-23Gasser, Les, A-14

Gasser, Les, A-15Geffner, Hector, A-16Geiger, Dan, A-19Generalization with Taxonomic Information,

A-24Generalized Shape Autocorrelation, A-18Generalized Theory of Model Preference, The,

A-22Genetic Algorithms and Classifier Systems, ,

A-19Gervasio, Melinda T., A-20Getting Serious About Parsing Plans: A

Grammatical Analysis of Plan Recognition,A-21

Giangrandi, Paolo, A-23Ginsberg, Allen, A-15Ginsberg, Matthew L., A-17, A-17Goel, Suraj, A-15Goldberg, David E., A-19Goldszmidt, Moisés, A-20Goodman, Marc, , A-22Gottlob, Georg, A-24Grosz, Barbara J., A-21Gruber, Thomas, A-14Guerreiro, Ramiro A. de T., A-19Guha, Ramanathan V., A-22Hadley, Robert F., A-23Hall, Lawrence O., A-16Halpern, Joseph Y., A-18Halpin, Michael P., A-17Hammond, Kristian J., A-15Hamscher, Walter, A-14Hanks, Steve, A-23Harada, Minoru, A-22Harris, Larry, A-6, A-20Harvey, William D., A-17Haussler, David, A-16Hayes, Philip J., A-21Heindel, Troy A., A-5, A-19Hellerstein, Joseph, A-22Hemerly, Andrea S., A-19Henry, Neil, A-21Hierarchical Planner that Generates Its Own

Hierarchies, A, A-20Hierarchical Protocol for Coordinating Mulita-

gent Behaviors, A,A-15Hinton, Geoffrey, A-14Hirschberg, Julia, A-16Hirst, Graeme, A-14Holcomb, Tim, A-3, A-7, A-19, A-22Hollan. James D., A-18Horty, John F., A-19Hölldobler, Steffen, A-23Hsia, Yen-Teh, A-18Huhns, Michael N., A-17Hussam, Abul, A-22Hybrid Connectionist Symbolic Learning Sys-

tem, A, A-16Hybrid Framework for Representing Uncertain

Knowledge, A, A-20Hyun, Seungseok, A-24Igarashi, Zenichi, A-22Incorporating Default Inferences Into Plan

Recognition, A-21Incremental Approximate Planning, A-19

Incremental Non-Backtracking Focusing: APolynomially Bounded GeneralizationAlgorithm for Version Spaces, A-24

Indexical Knowledge in Robot Plans, A-20Inductive Learning in a Mixed Paradigm Set-

ting, A-20Inductive Learning in Probabilistic Domain,

A-19Inductive Synthesis of Equational Programs,

A-18Inspector: An Expert System for Monitoring

Worldwide Trading Activities in ForeignExchange, A-21

Integrating Natural Language Processing andKnowledge Based Processing, A-20

Integrating Planning and Acting in a Case-Based Framework, A-15

Integrating, Execution, Planning, and Learn-ing in Soar for External Environments, A-23

Intelligent Database Interface: Integrating AIand Database Systems, The, A-23

Interaction-Based Invention: Designing NovelDevices from First Principles, A-21

Introducing the Tileworld: ExperimentallyEvaluating Agent Architectures, A-19

Introduction to Hypertext and Hypermedia, A-14

Irani, Keki B., A-20Ishida, Toru, A-15Issar, Sunil, A-17It’s Not My Default: The Complexity of Mem-

bership Problems in Restricted Proposition-al Default Logics, A-21

Iterative Broadening, A-17Iwayama, Makoto, A-15Jambor, Roger A-2, A-15Joe, Gene, A-17Johnson, Bruce B., A-7, A-22Johnson, P. E., A-17Johnston, Mark D., A-22Jones, Karen Sparck, A-18Jordan, Michael, A-14Junker, Ulrich, A-16Kabanza, Froduald, A-19Kalé, L. V., A-17Kambhampati, Subbarao, A-19, A-21Katagiri, Yasuhiro, A-16Kedar, Smadar, , A-19Keller, E. Robert, A-6, A-20Kitagawa, Yashhiro, A-22Kitano, Hiroaki, A-16Klahr, Philip, A-21Knoblock, Craig A., A-18Knowledge Level and Inductive Uses of

Chunking (EBL), A-24Kolodner, Janet, A-14Konolige, Kurt, A-16Koo, Sang Hoe, A-15Kosanovich, Karlene A., A-5, A-20Koseki, Yoshiyuki, A-19Kotas, Paul, A-21Koton, Phyllis, A-15Kounalis, Emmanuel, A-18Koza, John R., A-19Kramer, Glenn A., A-17Kraus, Sarit, A-22

A–26 INDEX

Kriegman, David J., A-18Krulwich, Bruce, A-24Kuipers, B., A-22Kuipers, Benjamin, A-19Laffey, Thomas, A-14Laird, Philip, A-22, A-24Lakemeyer, Gerhard, A-16Le Cun, Yann, A-14Learning Abstraction Hierarchies for Problem

Solving, A-18Learning Causal Trees from Dependence

Information, A-19Learning from Textbook Knowledge: A Case

Study, A-15Learning General Completable Reactive

Plans, A-20Learning Search Control for Constraint-

Based Scheduling, A-18Learning to Coordinate Behaviors, A-22Lee, Wing, A-15Lejter, Moises, A-24Lemke, Andreas C., A-23, A-21Lenzerini, Maurizio, A-21Lespérance, Yves, A-20Letovsky, Stanley, A-23Levesque, Hector J., A-16, A-21, A-20Levine, John M., A-16Levine, Rich A-2, A-15Ligozat, Gérard, A-17Liu, Yanxi, A-20Liu, Zheng-Yang, A-15Lochbaum, Karen E., A-21Lofaso, Bernie, A-24Logical Task Modeling for Man-Machine

Dialogue, A-16LOGnets: A Hybrid Graph Spatial Represen-

tation for Robot Navigation, A-24Loofbourrow, Tod, A-14Looking For the AI in Software Engineering:

An Applications Perspective, A-16MacIntosh, D. J., A-17Mackworth, Alan K., A-21Maes, Pattie, A-22Magerman, David M., A-20Mahler, Ed, A-6, A-22Maintaining Consistency in a Stratified Pro-

duction System Program, A-16Malkin, Peter K., A-24Mapping and Retrieval During Plan Reuse: A

Validation Structure Based Approach, A-19Marcus, Mitchell P., A-20Marks, Joseph, A-17Massively Parallel AI, A-23Matheus, Christopher John, A-20Matthews, C. E., A-17Maximum Entropy Approach to Nonmono-

tonic Reasoning, A, A-20McAllester, David, A-20McCall, Raymond, A-23McCune, William, A-19McDermott, John, A-16McKay, Don, A-2, A-15, A-23McKeown, Kathleen R., A-18Mechanizing Inductive Reasoning, A-18Meiri, Itay, A-18Method of Calculating the Measure of Salience

in Understanding Metaphors, A, A-15

Meyer, R.A., A-17Mick, Don A-2, A-15Miller, James, A-15Minton, Steven, , A-19, A-22Minton, Steven, A-22Miranker, Daniel P., A-24Mitchell, Tom M., A-24Mittal , Sanjay, A-22Mittal, Sanjay, A-14Model-Based Diagnosis of Planning Failures,

A-24Model-Based Diagnosis, A-14Models of Plans to Support Communication:

An Initial Report, A-21Mohan, Rakesh, A-18Moldovan, Dan, A-15Montgomery, Thomas A.,A-15Mooney, Raymond J., A-24, A-15Moore, Johanna D., A-17Moore, Robert, A-6, A-20Morgenstern, Leora, A-16Morris, Paul, A-20, A-23Mostow, Jack, A-18Mukerjee, Amitabha, A-17Mullarkey, Peter W., A-17Murnane, Didi, A-5, A-20Murray, William R., A-23Murtagh, Niall, A-17Musen, Mark, A-14Myers, Karen L., A-18Myths and Legends in Learning Classifica-

tion Rules, A-20Nadel, Bernard, A-14, A-18Nakakuki, Yoichiro, A-19National Dispatcher Router: A Multi-

Paradigm Based Scheduling Advisor…, A-22

Natural Language Interfaces, A-14Negotiation and Conflict Resolution in Non-

Cooperative Domains, A-17Nejdl, Wolfgang, A-24Nelson, Paul, A-14Ng, Hwee Tou, A-24Nielsen, Jakob, A-14Nonmonotonicity and the Scope of Reason-

ing, A-22Noordewier, Michiel O., A-16Novak Jr., Gordon S., A-17Nunes, José H. T., A-16O’Connor, Gregory M., A-5, A-20O’Hare, Anthony, A-23Obtaining Quantitative Predictions from

Monotone Relationships, A-22Okuide, Satoshi, A-22Okumura, Manabu, A-20On Acting Together, A-16On Analytical and Similarity-Based Classifi-

cation, A-15On the Complexity of Monotonic Inheritance

with Roles, A-19On the Expressiveness of Networks with Hid-

den Variables, A-21On the Performance of Lazy Matching in Pro-

duction Systems, A-24On the Role of Coherence in Abductive Expla-

nation, A-24

Operationality Criteria for Recursive Predi-cates, A-23

Optimally Efficient Limited Inference System,An, A-21

Organizational Approach to Adaptive Produc-tion Systems, An, A-15

Ourston, Dirk, A-15Paek, Eunok, A-16Page, C. David, A-24Palmer, Martha, A-20Parallel Asynchronous Distributed Production

System, A, A-15Parametric Engineering Design Using Con-

straint-Based Reasoning, A-17Park, Young-Tack, A-17Parsing a Natural Language Using Mutual

Information Statistics, A-20Passonneau, Rebecca, A-20Path-Focused Duplication: A Search Proce-

dure for General Matings, A-17Paz, Azaria, A-19Pearl, Judea, A-18, A-20, A-19Perlis, Donald, A-22Peterson, Gerald E., A-19Philips, Andrew B., A-22Physical Impossibility Instead of Fault Mod-

els, A-24Pickering, Cynthia, A-14Pinchover, Eli, A-18Pointing: A Way Toward Explanation Dia-

logue, A-17Pollack, Martha E., A-19Ponce, Jean, A-18Popplestone, Robin J., A-20Practical Temporal Projection, A-23PRAGMA—A Flexible Bidirectional Dialogue

System, A-16Prerau, David, A-14Principled Approach to Reasoning About the

Specificity of Rules, A, A-24Principles and Practice of Knowledge Acquisi-

tion, A-14Prism: A Case-Based Telex Classifier, A-22Probabilistic Interpretation for Lazy Non-

monotonic Reasoning, A, A-20Probabilistic Semantics for Cost Based

Abduction, A-19Probabilities that Imply Certainties, A-20Probably Approximately Correct Learning, A-

16Proven Domain-Independent Scientific Func-

tion-Finding Algorithm, A, A-22QPC: A Compiler from Physical Models into

Qualitative Dualitative Differential Equa-tions, A-22

Qualitative Model for Space, A, A-A-22Qualitative Reasoning About Physical Sys-

tems, , A-19Qualitative Reasoning with Microscopic The-

ories, A-15Rahaim, Jack A-2, A-18Rajamoney, Shankar A., A-15Ramesh, Bhashyam, A-21Randazza, Margaret, A-24Randy, Raynor, A-6Rappaport, Alain, A-6, A-20Raschid, Louiqa, A-16

AAAI ‘90 A–27

Rationality and Its Roles in Reasoning, A-17Real-Time Knowledge-Based Systems, A-14Reasoning about Qualitative Temporal Infor-

mation, A-17Reasoning and Acting in Real Time, A-23Redmond, Michael, A-15Refinement of Approximate Domain Theories

by Knowledge-Based Neural Networks, A-16

Reinstein, Harry C., A-6, A-20Reiter, Ehud, A-17Reiter, Raymond, A-21Rendell, Larry, A-19Representation of Defaults in Cyc, The, A-22Riesbeck, Chris, A-14Ringuette, Marc, A-19Rissland, Edwina L., A-20Rogalski, Mike, A-A-18, A-19Romaniuk, Steve G., A-16Rosenbloom, Paul S., A-24, A-24, A-24, A-

23Rosenschein, Jeffrey S., A-17Rosenschein, Jeffrey, A-14Rosenschein, Stanley J., A-23Rothstein, Janet, A-22Rusinowitch, Michaël, A-18Rychlik, Piotr, A-22Sacerdoti, Earl, A-14Sacks, Elisha, A-19Sadek, M. D., A-16Saffiotti, Alessandro, A-20Saletore, Vikram A., A-17Satoh, Ken, A-20Saxena, Sharad, A-16Schaffer, Cullen, A-22Schelm, Roger, A-3, A-19Schlimmer, Jeffrey C., A-22Schmiedel, Albrecht, A-19Schmolze, James G., A-15Schoen, Paul D., A-5, A-20Scovell, Keith, A-3, A-7, A-19, A-22Search Lessons Learned from Crossword Puz-

zles, A-17Self-Explanatory Simulations: An Integration

of Qualitative and Quantitative Knowl-edge, A-22

Selman, Bart, A-21Shachter, Ross D., A-18Shastri, Lokendra, A-21Shavlik, Jude W., A-16Shelm, Roger, A-3, A19Shen, Wei-Min, A-24Shifting Ontological Perspectives in Reason-

ing About Physical Systems, A-15Shimony, Solomon E., A-19Shimura, Masamichi, A-17Shoham, Yoav, A-19Shrobe, Howard, A-21, A-22Shvaytser, Haim, A-20Sidner, Candace L., A-18Siegle, Greg, A-23Silverman, Steven, A-21Simmons, Robert F., A-20Simoudis, Evangelos, A-15Six Companies Reveal Results of Expert Sys-

tems Successes, A-2Skalak, David B., A-20

Skolem Functions and Equality in AutomatedDeduction, A-19

Smith, Benjamin D., A-24Smith, Ted A-2, A-18Soloway, Elliot, A-16Solving Geometric Constraint Systems, A-17Solving Large-Scale Constraint-Satisfaction

and Scheduling Problems Using a HeuristicRepair Method,A-22

Solving Term Inequalities, A-19Some Applications of Graph Bandwidth to

Constraint Satisfaction Problems, A-18Sridharan, N. S., A-14Sriram, Duvvuru, A-19Stillman, Jonathan, A-21Strauss, Peter, A-14STRIPS Assumption for Planning Under

Uncertainty, The, A-21Structure of Perspectivity; A Case of Japanese

Reflexive Pronoun “zibun,” A-16Structured Connectionist Unification Algo-

rithm, A, A-23Subramanian, Devika, A-23Swartout, William R., A-17Symbolic Probabilistic Inference in Belief Net-

works, A-18Symmetry Constraint Inference in Assembly

Planning: Automatic Assembly Configura-tion Specification, A-20

Synthesis of Reactive Plans for Multi-PathEnvironments, A-19

Tambe, Milind, A-24Tan, Ming, A-22Tanaka, Hozumi, A-15, A-20Tanaka, Midori, A-19Tasso, Carlo, A-23Techies Versus the Nontechies: Today’s Two

Cultures, The, A-16Temporal Terminological Logic, A, A-19Tenenberg, Josh D., A-21Terminological Cycles in KL-ONE-based

Knowledge Representation Languages, A-19

Terveen, Loren G., A-21The Future of Knowledge Representation, A-

14Theory of Plan Modification, A, A-21Theory Reduction, Theory Revision, and

Retranslation, A-15Thomason, Richmond H., A-19Tieman, Larry A-2, A-15Tokunaga, Takenobu, A-15Torrance, Mark C., A-17Touretzky, David, A-14Towards a System Architecture Supporting

Contextualized Learning, A-23Towards Incremental Disambiguation with a

Generalized Discrimination Network, A-20Towell, Geoffrey G., A-16Tree Decomposition with Applications to

Constraint Processing, A-18Truly Parallel Understanding of Text, A-20Truth Maintenance Systems, A-14Truth Maintenance, A-20Two Case Studies in Cost-Sensitive Concept

Acquisition, A-22Two Views of Belief: Belief as Generalized

Probability and Belief as Evidence, A-18Understanding Natural Language with Dia-

grams, A-17User Modeling and User Interfaces, A-18User Models and User Interfaces: A Case for

Domain Models, Task Models, and Tai-lorability, A-18

Utgoff, Paul E., A-16Utility of EBL in Recursive Domain Theories,

The, A-23Validated Retrieval in Case-Based Reasoning,

A-15van Beek, Peter, A-17Very Fast Decision Table Execution of Propo-

sitional Expert Systems, A-24Vilain, Marc, A-21, A-15Volovik, D., A-17Waldman, Steven, A-21Waltz, David L., A-23Watanabe, Larry, A-19Weak Representations of Interval Algebras, A-

17Weinstein, Steven, A-21Weir, Carl, A-20Weld, Daniel S., A-15Wellman, Michael P., A-21What Should Be Minimized in a Decision

Tree?, A-20What Users Are Buying Today in AI and

Why: Reports from Management, A-3What’s In a User?, A-18White, John L., A-19Why and How Corporate America Is Integrat-

ing AI into Data Processing, A-2, A-18Why Is It So Hard Moving AI to Reality?, A-

14Why PRODIGY/EBL Works, A-23Wilkins, David C., A-17Williams, Brian, A-21Williams, Chuck, A-2, A-15Wise, Dave, A-4, A-19Woods, John A-2, A-15Wozniak, Stanley, A-7, A-22Wroblewski, David A., A-21Wu, Thomas D., A-24Yang, Qiang, A-22, A-21Yee, Richard C., A-16Yeh, Alexander, A-15Yen, John, A-24Yokoo, Makoto, A-15Yu, Dorothy, A-5, A-20Yu, Yeong-Ho, A-20Zabih, Ramin, A-18Zhu, Jieming, A-16Zlotkin, Gilad, A-17Zualkernan, I. A., A-17Zweben, Monte, A-18Zytkow, Jan M., A-22

A–28 INDEX

Program & Schedule · involved in integrating expert system technology in commercial environments...

Documents

Transcript of Program & Schedule · involved in integrating expert system technology in commercial environments...