tr^c

A-92 . It is being mailed to you in >S. UMBC.EDU. For further information, _/l-1013 or at [email protected].

ADVANCE PROGRAM

The Eighth lEEE Conference on for APPLICATIONS

March 2-6, 1992 Doubletree Hotel Monterey, California

CONFERENCE COMMITTEE

General Chair: Timothy Finin, University of Maryland, Baltimore County Program Chair: Jan Aikins, Aion Corporation Tutorial Chair: Daniel 0 Leary, University of Southern California Workshop Chair: Donald McKay, Unisys Defense Systems Publicity Chairs: Curt Hall and Paul Harmon, Intelligent Software Strategies newsletter Local Arrangements Chair: Bob Engelmore, Stanford University

Program Committee: Chidanand Apte, IBM Jim Bennett, Amdahl Ron Brachman, AT&T Bell Labs Elizabeth Byrnes, Bankers Trust Joe Carter, Andersen Consulting Vasant Dhar, New York University Lee Erman, Cimflex Teknowledge Corporation Richard Gabriel, Lucid Inc. Phil Hayes, Carnegie Group, Inc. Se June Hong, IBM Gary S. Kahn, A. C. Nielsen Bernadette Kowalski Minton, Aion Corporation William Mark, Lockheed IT£C Use , ; L---**'

Date: Tue, 7 Jan 92 16:49:12— EST From: [email protected] To: [email protected]

This is the advance program for CAIA-92. It is being mailed to you in response to a message sent to [email protected]. For further information, contact the lEEE CS Office at 202-371-1013 or Tim Finin at [email protected].

ADVANCE PROGRAM

The Eighth lEEE Conference on ARTIFICIAL INTELLIGENCE for APPLICATIONS

March 2-6, 1992 Doubletree Hotel Monterey, California

CONFERENCE COMMITTEE

General Chair: Timothy Finin, University of Maryland, Baltimore County Program Chair: Jan Aikins, Aion Corporation Tutorial Chair: Daniel 0 Leary, University of Southern California Workshop Chair: Donald McKay, Unisys Defense Systems Publicity Chairs: Curt Hall and Paul Harmon, Intelligent Software Strategies newsletter Local Arrangements Chair: Bob Engelmore, Stanford University

Program Committee: Chidanand Apte, IBM Jim Bennett, Amdahl Ron Brachman, AT&T Bell Labs Elizabeth Byrnes, Bankers Trust Joe Carter, Andersen Consulting Vasant Dhar, New York University Lee Erman, Cimflex Teknowledge Corporation Richard Gabriel, Lucid Inc. Phil Hayes, Carnegie Group, Inc. Se June Hong, IBM Gary S. Kahn, A. C. Nielsen Bernadette Kowalski Minton, Aion Corporation William Mark, Lockheed Brian McCune, Booz, Allen & Hamilton Inc. Steven Minton, NASA Ames Research Center Fumio Mizoguchi, Tokyo Science University Ramesh Patil, USC/Information Sciences Institute Earl Sacerdoti, The Copernican Group Lokendra Shastri, University of Pennsylvania Howard E. Shrobe, Symbolics Inc. , Thinking Machines Corp. Mike Wellman, Wright-Patterson AFB Mike Williams, IntelliCorp

CONFERENCE OVERVIEW

During the last seven years, the lEEE Computer Society's Conference on Artificial Intelligence for Applications has established itself as a key conference for business and technical people who want to keep up on the fast-changing world of commercial AI . This year the eighth annual CAIA 92 will continue that tradition by providing a provocative mix of introductory, intermediate and advanced talks, tutorials, workshops, and panels. This year's conference will feature a special series of talks and panels on the various knowledge representation standardization efforts that are underway in a variety of organizations. The plenary talks will be:

AI Applications at NASA: Hindsight and Foresight, Dr. Melvin Montemerlo, NASA

Software Patents and Copyrights: How Did We Get Into This Mess? And How Do We Get Out? Dr. Randall Davis, MIT

The Calculus of Fuzzy If-Then Rules and Its Application, Dr. Lotfi Zadeh, University of California

CAIA-92 TECHNICAL PROGRAM - Wednesday, March 4 8:45 - 9:00 am CHAIRPERSON'S WELCOME ADDRESS: Timothy Finin, University of Maryland and Unisys Defense Systems

9:00 - 10:00 am KEYNOTE TALK:

AI Applications at NASA Hindsight and Foresight, Dr. Melvin Montemerlo, Program Manager for Automation and Robotics, NASA Headquarters Chair: Timothy Finin, University of Maryland and Unisys Defense Systems

10:30 am - 12:10 pm Concurrent Sessions

PAPER SESSION: INTEGRATING RULE-BASED SYSTEMS Session Chair: Chidanand Apte, IBM T. J. Watson Research Center

SPOTLIGHT: A Data Explanation System, Tej Anand and Gary Kahn

GRAIL: An Integrated Artificial Intelligence System for Gene Recognition and Interpretation, X. Guan, R. C. Mann, R. Einstein, R. Mural and E. Überbacher

Intelligent Document Generation System for Construction Planning, Takamune Kita zawa, Mitsutoshi Hayata, Motoo Matsuda and Shigehito Kaji

Pitch Expert An Engineered Collection of Specialized Knowledge Structures, Allen Kowalski, Daniel Gauvin, Oliver Vadas and Yves Larin

PAPER SESSION: ENABLING TECHNOLOGIES Session Chair: Lokendra Shastri, University of Pennsylvania

Generalized Quicksearch for Expert Systems, Maurice Karnaugh

Constraint Propagation Over Ordered Domains, Bai Li, E. Burke, D. Elliman and P . Ford

Uncertainty Representation in Practical Decision Support Systems for the Field Service of Large Systems, Narendra Gupta

Representation and Control of Knowledge Bases for Support of Multiple Tasks, Young-Tack Park and David C. Wilkins

INVITED TALK:

Standards and Standardization: Assured Incompatibility? Carl Cargill, Digital Equipment Corp.

INVITED TALK:

Overview of Standards Activities Earl Sacerdoti, The Copernican Group

12:10 - 1:30 pm Lunch and Sun 1:30 - 3:30 pm Concurrent Sessions

Action Selection in Interactive Model-Based Diagnosis, Jiah-Shing Chen and Sarg ur N. Srihari

Automated Diagnosis for the Time of Flight Scintillation Array: Development of a Structural and Behavioral Reasoning System, Robert K. Paasch and Alice M. Agogino

A Multi-level Diagnosis Methodology for Complex Systems, Xudong Yu and Gautam Biswas

An Expert System to Perform Functional Diagnosis of a Bus Subsystem, Chris Preist, Daryl Allred and Ajay Gupta

Combining Constraint Satisfaction and Local Improvement Algorithms to Construct Anaesthetists Rotas, Barbara M. Smith and Scan Bennett

Balancing Antagonistic Time and Resource Utilization Constraints in Over-Subscr ibed Scheduling Problems, Stephen F. Smith and Dhiraj K. Pathak

An Assumption-Based Combinatorial Optimization System, Hirotaka Hara, Nobuhiro Yugami and Hiroyuki Yoshida

A Constraint Satisfaction Approach to the Resolution of Uncertainty in Image Interpretation, Peter B. Cullen, Jonathan J. Hull and Sargur N. Srihari

INVITED TALK:

Task Specific Architectures in Knowledge-Based Systems: Progress and Issues B. Chandrasekaran, Ohio State University

PANEL: The Proper Role of Standards in AI Panel Chair: Mike Williams, IntelliCorp

4:00 - 5:45 pm Concurrent Sessions 6:30 - 8:00 pm RECEPTION

CAIA-92 TECHNICAL PROGRAM - Thursday, March 5 9:00 - 10:00 am KEYNOTE TALK:

Software Patents and Copyrights How Did We Get Into this Mess? How Do We Get Out? Dr. Randall Davis, Professor of and Associate Director of the AI Lab, Massachusetts Institute of Technology

Chair: Jan Aikins, Aion Corporation

10:30 am - 12:10 pm Concurrent Sessions

PAPER SESSION: KNOWLEDGE ACQUISITION AND LEARNING I Session Chair: Phil Hayes, Carnegie Group, Inc.

KITSS A Functional Software Testing System Using a Hybrid Domain Model, Uwe Nonnenmann and John K. Eddy

An Expert System for the Generation of Digital Frequency Discriminator Multiplexes, Douglas E. Landon

Learning Approximate Diagnosis, Yousri El Fattah and Paul O Rorke

Knowledge Acquisition Support Environment (KASE) , Padma D. Reddy and Dick B. Simmons

PAPER SESSION: INTELLIGENT INTERFACES AND NATURAL LANGUAGE PROCESSING Session Chair: Dave Waltz, Thinking Machines Corp.

Integrated Software Environments, Christopher Landauer

Intelligent Assistance for the Communication of Information in Large Organizations, Robert Kass and Irene Stadnyk

Explanation Based on Contexts, Pertti Huuskonen and Anneli Korteniemi

Natural Language Processing Environments Based on a Multilingual Workstation, Richard F. Walters and Cvi Zhang PANEL :

As CASE Approaches AI Panel Chair: Mike Williams, IntelliCorp

12:10 - 1:30 pm Lunch and Sun

1:30 - 3:30 pm Concurrent Sessions

PAPER SESSION: KNOWLEDGE ACQUISITION AND LEARNING II Session Chair: Daniel O Leary, University of Southern California

Using Machine Learning to Understand Manufacturing Control Issues, Bradley L. Whitehall, Bill Fulkerson, James Hall and Stephen C-Y. Lv

Building Expert Systems by Training with Automatic Neural Network Generating Ability, Hahn-Ming Lee and Ching-Chi Hsu

Learning to Predict DNA Hydrations Patterns, Dawn Cohen, Bohdan Schneider, Helen Berman and Casimir Kulikowski

KEDS : A Knowledge-based Equation Discovery System for Engineering Problems, R. Bharat Rao and Stephen C-Y. Lv

Learning and Representing Concepts with Graded Structure, Jianping Zhang

PAPER SESSION: CASE-BASED REASONING, GENETIC ALGORITHMS AND FUZZY LOGIC Session Chair: Se June Hong, IBM T. J. Watson Research Center

Extending the Indexing Vocabulary of Case-Based Reasoning with Task Specific Features, Olivier Fischer, Jack W. Smith Jr. and Philip Smith

Use of Multiple Cases in Case-Based Design, Toshikazu Tanaka, Masakazu Hattori and Naomichi Sueda

MICAH: A Genetic Algorithm System for Multi-Commodity Network Problems, Sam R. Thangiah and Kendall E. Nygard

IDA-NET: An Intelligent Decision Aid for Battlefield Communications Network Configuration, W. D. Potter, R. Pitts, P. Gillis, J. Young and J. Caramadre

Quantitative/Fuzzy Approach to Document Recognition, Hiroko Fujihara and Elmamoun Babiker Panels:

PANEL :

Extracting Information from Text Panel Chair: Phil Hayes, Carnegie Group, Inc.

PANEL :

Data Mining and Information Creation Panel Chair: Vasant Dhar, New York University

4:00 - 5:30 pm PANEL: Case-Based Reasoning — Ready for Prime Time? Panel Chair: Esther Dyson, EDventure Holdings

CAIA-92 TECHNICAL PROGRAM - Friday, March 6 9:00 - 10:00 am KEYNOTE TALK:

The Calculus of Fuzzy If-Then Rules and Its Applications, Professor Lotfi Zadeh, Emeritus, University of California, Berkeley Chair: David Waltz, Thinking Machines Corp.

10:30 am - 12:10 pm Concurrent Sessions

PAPER SESSION: DESIGN Session Chair: Bob Engelmore, Knowledge Systems Laboratory, Stanford Universit

Toward Theory of Group Design Activities: Emphasizing Engineering Design, Youngwhan Lee

Design, Evaluation and Redesign Handling the Ad-hocness of a Knowledge-based Design System, H. Liv, C. D. Rowles and W. Wen

A Novel Approach to Knowledge-Based Communication System Design, Michael Edrich

The Synthesis of Simulation Programs Used in CCD Design, Takaaki Tatsumi, Takashi Usuki and Yoshiaki Kagawa PAPER SESSION: DISTRIBUTED PROBLEM SOLVING AND REAL-TIME SYSTEMS Session Chair: Timothy Finin, University of Maryland and Unisys Defense Systen

A Transaction Model for Multiagent Production Systems, Toru Ishida

Distributed Big Brother, Young-pa So and Edmund H. Durfee

MIP: A Real-Time Expert System, Xavier Alaman, Soledad Romero, Carlos Aguirre, Pedro Serrahima, Rafael Munoz, Vincente Lopez, Jose Dorronsoro, Erik de Pablo

Integrating Model-Based and Heuristic Features in a Real-Time Expert System for Power Distribution Networks, Monika Pfau and Wolfgang Nejdl

A Fully Integrated Real-Time and Multi-Tasking Knowledge-Based System: Application to an On-Board Diagnostic System, Gilles Gerlinger and Pierre Morizet-Mahoudeaux

PANELS :

Scheduling Technologies Panel Chair: Don Rosenthal, Three Sigma Systems

Integrating Information Technologies in AI Panel Chair: Neena Buck, New Science Associates

12:10 pm Conference Ends

CAIA-92 - TUTORIAL PROGRAM - Monday, March 2 8:30 am - 12:00 noon

TUTORIAL 1: CASE, Knowledge-Based Systems and Object-Oriented Programming LECTURERS: Jan Aikins, Aion Corporation, and Paul Harmon, Object-Oriented Strategies newsletter

This tutorial will introduce participants to the concepts underlying Computer-Aided Software Engineering (CASE) , object-oriented programming (OOP) and knowledge-based systems. We will consider the development of the CASE market in the U .S. and the role that OOP and knowledge-based systems are playing in the development of that market

We will consider popular CASE tools, some object-oriented tools and two or three expert system tools that are each playing a role in the U.S. CASE market to illustrate the advantages and problems associated with each approach to system development. We will also discuss the experiences of some U.S. companies that have used both conventional and intelligent CASE products.

This tutorial is designed to provide participants with an overview of the current CASE market in the U.S. and with an understanding of the evolving role that object-oriented and expert system techniques are playing within the software development market.

TUTORIAL 2: Planning: Techniques and Applications LECTURERS: Steven Minton and Mark Drummond, NASA Ames Research Center

Planning has been clearly recognized throughout the history of AI as an import ant and difficult problem. Many practical problems exist which can be solved by the appropriate application of AI planning techniques, often in conjunction with techniques from related fields such as operations research, control theory, and expert systems.

The tutorial begins with a motivation of the planning problem and establishes simple notation and terminology. Issues such as operator and plan representation, commitment strategies, abstraction, and search are covered. We then turn to a discussion of the methods by which a plan might be executed in an external environment; this leads naturally to a discussion of modern reactive systems. Following this, we present a section on learning techniques and their application to planning problems. Related techniques are covered next, including operations research and control theory. The last section presents some practical applications of planning systems over the last decade. We close with a summary of all major issues raised in the tutorial, and provide an opportunity for tutorial attendees to ask questions.

The tutorial is intended for those who develop, use, or manage knowledge-based systems in research or industry, where there are problems which might require the use of AI planning techniques. It is also intended for active researchers in AI who want to improve their understanding of planning techniques or want to find out about recent progress in the field. It assumes a familiarity with basic AI concepts at the level of an introductory textbook.

TUTORIAL 3: Case-Based Reasoning LECTURER: Evangelos Simoudis, Lockheed AI Center

Case-Based Reasoning (CBR) has recently found its way out of the university laboratories and been applied to a variety of real-life tasks including diagnosis and design. In this tutorial we will provide an introduction to CBR, examine CBR in relation to other reasoning methods such as rule-based and model-based re asoning, and describe the process of implementing a case-based reasoning system . In the introduction we will describe the major components of a CBR system, the reasoning cycle performed by such a system, and provide a generic architectur e for case-based reasoning. We will also investigate issues of case base organization, indexing, adaptation, and learning. We will describe the process of imp lementing a CBR system by examining the type of algorithms that need to be implemented for such a system, the domain-specific knowledge that needs to be acqui red, and methods for transforming the contents of existing databases to case bases. Various aspects of the implementation process will be illustrated with exa mples from real-life CBR systems from academia and industry. Finally, we will explore possible tasks that can be solved through the use of case-based reasonin g, and identify advantages and disadvantages of the method by evaluating existing systems .

12:00 noon - 1:30 pm: LUNCH 1:30 pm - 5:00 pm

TUTORIAL 4: Project Management and Development of Expert Systems in Business and Industry LECTURER: Dr. David S. Prerau, GTE Laboratories Inc.

This tutorial provides proven, practical techniques for project management and development of expert systems in business and industry. The objective of the tutorial is to prepare the attendees to manage expert systems effectively from the initial steps of selecting an application through the development process, to the transfer and deployment of the system. The discussion will be supported by examples drawn from case studies.

The tutorial will present an overview of expert systems, including a discussion of the kinds of tasks for which they are suitable and the types of benefits they can provide. The stages in a typical expert system development project will be described and discussed, as well as the major planning and management issue s involved.

The tutorial will focus on key issues in expert system development and present practical approaches to address them. First, there will be a discussion of the expert system team and the selection of domain experts. Next, the tutorial will investigate the evaluation of the application domain for expert systems and will present a set of technical and nontechnical criteria to guide the evaluatio n. Then knowledge acquisition and important technical and logistical issues wil 1 be discussed. Pragmatic issues that impinge on the success of an expert system project will be highlighted.

This tutorial is intended for managers, professionals, programmers, and software engineers in business and industry who are interested in applying and integr ating expert systems technology into their current work. The tutorial is direct ed primarily to persons who do not have significant experience in developing ex pert systems, and it does not presuppose technical knowledge of expert system techniques.

TUTORIAL 5: Knowledge Acquisition Techniques LECTURERS: Jan Clayton and Carli Scott, Expert Support Inc.

Knowledge acquisition is one of the greatest challenges in building expert systems. In the past, knowledge engineers learned how to conduct productive knowledge acquisition sessions from trial and error. In recent years the knowledge engineering community has developed more mature methodologies for developing expert systems. This tutorial presents an introduction to knowledge-acquisition methodology .

The tutorial focuses on the relationship between knowledge acquisition and the development of expert systems. The most common knowledge-acquisition activitie s will be discussed, along with how those activities change as the project progresses. In addition, the presenters will describe guidelines, techniques, and procedures that have proven successful in developing applications . The tutorial presenters will focus on the practice of knowledge acquisition and on what questions to ask. In addition, the tutorial will cover problems that knowledge engineers are likely to encounter, how to avoid these problems, and how to correct them when they do occur.

The tutorial will not cover: Automated knowledge acquisition; Knowledge acquisition as an interviewing activity; Knowledge-acquisition techniques as they relate to specific domain types .

The tutorial is aimed at new expert systems developers and those with a modest amount of experience. We assume that audience members have a basic understanding of expert-system concepts and terminology. In addition, we expect that audience members will be attending the tutorial to understand how knowledge acquisition relates to the development of expert-system applications . This tutorial is not appropriate for the research community.

TUTORIAL 6: Practice Cannot Be Reduced to Theory — An Introduction to Social Science Perspectives on Software Design LECTURER: William J. Clancey, Institute for Research on Learning

This tutorial serves as a bridge between the language and approaches of software engineers, including expert systems programmers, and researchers from other disciplines who study technology, design, and the workplace.

The disciplines of human factors, ethnography, sociolinguistics, systems analysis, and knowledge engineering are now being synthesized into an integral metho dology for understanding how to design tools to help people work better together. The challenge is how to relate observations of human behavior, theories abou t knowledge and representations, and the tool design process. My approach to situated cognition provides a justification for social scientists rejection of a utomation as the only way of using computers in the work place. This is summarized by the aphorism, Practice cannot be reduced to theory.

The tutorial combines case studies and theoretical analysis, focusing on social science claims that human interactions in the workplace cannot be equated to models, that expert systems approaches to automating practice lead to rigidification and deskilling, and recent approaches to cognition that sharply contrast computer models with human knowledge and reasoning. An example analysis and design of an information processing system for a chemical manufacturing plant is w orked out in detail.

This tutorial is specifically designed for managers, researchers, and programmers who are generally familiar with knowledge engineering. Explanations of soci al science perspectives are basic and introductory; no previous experience is expected or required.

CAIA92 TUTORIAL PROGRAM - Tuesday, March 3

8:30 am - 12:00 noon

TUTORIAL 7: Developing Classification Rules (Trees) from Examples LECTURER: Se June Hong, IBM T. J. Watson Research Center

Classification rules or decision trees are used in many knowledge-based applications. Developing such rules directly from examples is an important way to acq uire knowledge. This tutorial aims to accomplish three things: present enough background and discussion to understand the issues (e.g., rules in learning pers pective, features, examples, ID3, AQ approaches); present several specific new algorithms for extracting essential features and developing near minimal set of rules (with emphasis on efficiency and runtime space comparable to the space required to hold the examples) ; and to discuss related issues including errors, sample size, and actual use of classification rules.

The audience should have some experience in developing algorithmic programs. Background in logic minimization will be helpful. This tutorial is for those motivated to develop rules from examples with symbolic feature values.

TUTORIAL 8: Knowledge-Based Production Management LECTURERS: Dr. Stephen F. Smith and Dr. Norman Sadeh, Carnegie Mellon Universi

A large number of efforts have sought to investigate and exploit the use of AI techniques in production management applications. This work has attempted to respond to the inadequacies of existing computer-based solutions and the limited impact that Operationsßesearch and Operations Management have had. Research in knowledge-based production management emphasizes solutions that match the requirements, characteristics and constraints of practical problems .

This tutorial presents an overview of the concepts, techniques and methodologies in knowledge-based production management, and provides an assessment of the current state of theory and practice. The sources of difficulty associated with the production management and control problem are reviewed, as are the requirements for the development of effective production management tools. Current production management practice is examined and the shortcomings of traditional approaches are identified. Opportunities provided by knowledge-based technologies contributing to effective implementation of modern manufacturing philosophies are discussed.

Issues relating to the injection of knowledge-based production management into actual factory operations are considered and a few successful applications are reviewed. The tutorial concludes with a discussion of important issues for further research and the challenges these issues present vis a vis the current state of the field.

This tutorial is aimed at both practitioners and researchers who are interested in applying knowledge-based techniques to practical production management problems.

TUTORIAL 9: Real-Time Knowledge-Based Systems LECTURER: Thomas J. Laffey, Talarian Corporation

A real-time system is one in which correctness depends not only on the results of the computation, but also on the time at which the results are produced. Real-time computer systems are being used in a growing number of applications ranging from small, simple controllers found in common household appliances to large, complex systems for industrial and military purposes. The complexity of these systems is increasing rapidly and this has caused considerable interest in using knowledge-based techniques for real-time applications.

This tutorial presents the current state-of-the-art in applying knowledge-based problem solving methods to complex real-time applications. Ongoing research in key theoretical areas will be discussed, including: (1) high performance and predictable response times, (2) time-constrained reasoning, and (3) focus of attention.

The tutorial will also survey what commercial AI products are available for real-time problem solving. We will conclude with a study of several applications in the space industry.

Individuals interested in knowledge-based problem solving methods to real-time problems will benefit from this tutorial. It will be useful to managers, AI professionals, and researchers who want to know what is commercially available and what problems need to be solved before we see the widespread use of real-time knowledge-based systems. This tutorial assumes familiarity with AI and knowledge-based systems .

12:00 noon - 1:30 pm: LUNCH 1:30 pm - 5:00 pm

TUTORIAL 10: Verification and Validation of Knowledge-Based Systems LECTURERS: Robert M. 0 Keefe, Rensselaer Polytechnic Institute, and Daniel E. O Leary, University of Southern California

Methods and techniques for verifying and validating knowledge-based systems are reviewed. Verification methods are presented including well known techniques for investigating the consistency and completeness of a rule base. The use of these techniques in verifying frame-based systems is also considered, and recent developments in automated tools for verification are presented.

The validation process is then discussed, particularly the establishment of criteria by which a system is judged, the need to maintain objectivity, and the concept of reliability. This is followed by a review of validation techniques for validating both the components of a system and the system as a whole, and includes examples of some useful statistical methods. Management of the verification and validation process is then considered.

For those whose verification and validation methods are ad hoc or poorly developed, the tutorial will provide the knowledge to implement and manage better procedures, resulting in knowledge-based systems that are more reliable and of hi gher quality.

This tutorial is intended for knowledge engineers, software engineers, AI researchers and others who develop or manage the development of knowledge-based systems. It is pitched at an intermediate level suitable for practicing professionals. It is assumed that participants are familiar with basic knowledge representation methods such as rules and frames, common methods for handling uncertainty such as certainty factors and Bayesian networks, and basic software engineering concepts such as testing procedures and life cycle models.

TUTORIAL 11: Behavior-Based Reactive Robotic Systems LECTURER: Ronald C. Arkin, Georgia Institute of Technology

Traditional approaches to planning and control using AI techniques for the navigation of mobile robots are generally based upon reasoning over abstract models of the world. These models are either generated from a priori knowledge or are derived from sensory information. Decision-making is based upon this abstracted representation of reality. It has been shown that there are several pitfalls associated with this approach.

Reactive control strategies have been created in response to the limitations of model-based planning and control techniques. Abstract models of the world are seschewed in favor of the immediate utilization of sensory data. Robot response is not mediated by a model but is directly invoked from one or more sensory sources.

In this tutorial we discuss the motivating influences for behavior-based robotic systems and its roots in psychology, ethology and neuroscience. This is then followed by a discussion of the appropriate role of sensing and representation and how perceptual algorithms can be designed and integrated by the practitioners. A survey of behavior-based robotic systems is then presented with a critical analysis of each. Finally, the role and design of hybrid reactive/hierarchical architectures is presented and illustrated with specific examples.

This tutorial is aimed at engineering and computer science professionals who are concerned with developing real-time sensory response in robotic plan execution. Attendees will obtain both an understanding of the new developments in reactive robotic systems and an understanding of the techniques and processes necessary to construct behavior-based robotic systems. TUTORIAL 12: Graphical User Interface Design LECTURER: Aaron Marcus, Aaron Marcus and Associates

Graphical user interfaces (GUIs) are crucial for the success of computer-based products. The tutorial will introduce terminology, principles, guidelines, and heuristics for using information-oriented, systematic graphic design in user interfaces, especially for the design of icons, control panels and dialogue boxes, and navigational devices.

Participants will learn practical principles that are immediately useful, become familiar with many existing techniques, and discover potential new research topics. They will observe and analyze techniques for making products and displays more intelligible, functional, aesthetic, and marketable. Extensively illustrated lectures and video examples will cover perceptual, conceptual, and communication issues in typography, symbol systems, color, spatial composition, animation, and sequencing. The principles, guidelines, and case studies will be relevant for all window manager paradigms such as Motif and Open Look, all application software, on all platforms, with all input/output devices. The course will emphasize metaphors, mental models, navigation in the model, appearance, and interaction issues as well as the design process.

Many factors will be considered for designed products that are easy to produce, sell, learn, use and maintain. These include the market, compatibility, industry standards, corporate identity, development environment, user-centered design and productivity.

This tutorial will give developers, designers, and users valuable insight into key issues and show how to achieve effective visual communication .

CAIA-92 TUTORIAL LECTURER BIOGRAPHIES

Jan Aikins is a founder and current President of Technology for AION, an expert system s tool vendor for IBM PC and mainframe tools. At AION she is responsible for assessing technological developments in Artificial Intelligence and related fields. Prior to joining AION, Dr. Aikins held research positions at both Hewlett-Packard and IBM where she gained extensive experience in creating expert systems for both scientific and commercial applications. Dr. Aikins received her Ph.D. from Stanford University in Artificial Intelligence in 1980.

Prof. Ronald C. Arkin is an Associate Professor in the College of Computing at the Georgia Institute of Technology. He received a Ph.D. in Computer Science from the University of Massachusetts, Amherst, in 1987. From 1977 until 1985 he was a member of the faculty of Hawthorne College in Antrim, N.H. Dr. Arkin has authored more than 40 papers in these areas . He is a senior member of the lEEE and a member of ACM and AAAI . He is included in several Who's Who publications including Who's Who in the World.

William J. Clancey is a senior research scientist at the Institute for Research on Learning in Palo Alto, California. He has been active in expert system research since he joined the MYCIN project in 1975. He received his Ph.D. in computer science from Stanford University in 1979. He is editor-in-chief of AAAI Press, a senior editor of Cognitive Science, co-editor of Interactive Learning Environments, and member of the editorial board of Artificial Intelligence. He was also a cofounder and on the board of directors of Teknowledge, Inc. and, more recently, ModernSoft, Inc. He edited (with E. H. Shortliffe) Readings in Medical Artificial Intelligence: The First Decade (Addison-Wesley, 1984), wrote Knowledge-Based Tutoring: The GUIDON Program (MIT Press, 1987), and also edited (with E. Soloway) Artificial Intelligence and Learning Environments (MIT Press, 1990) .

Jan Clayton has 10 years of experience in the field of artificial intelligence and expert-systems technology. Her research and commercial experience includes : Application development of small- and large-scale expert systems; Documentation and training development; and Expert-systems methodology development. She was a member of Stanford's Heuristic Programming project from 1978 to 1982. In 1982 she joined Teknowledge Inc. as a knowledge engineer. During her 7 years at Teknowledge, Ms. Clayton participated heavily in both application and training development. As such, she worked on several successfully fielded expert systems. In 1990, she became a founder of Expert Support Inc.

Mark Drummond received a Ph.D. in Artificial Intelligence from the University of Edinburgh in 1987. He worked for two years at the AI Applications Institute in Edinburgh as a consultant on planning and scheduling for various European companies and government organizations. Dr. Drummond is currently employed by Sterling Software, and works at the NASA Ames Research Center. Dr. Drummond is leader of the Entropy Reduction Engine project, a compact team of enthusiastic researchers, focusing on the integration of techniques from AI planning, scheduling, and discrete event control. Dr. Drummond is the author or co-author of more than 15 publications in the areas of planning and scheduling.

Paul Harmon is the editor/author of two newsletters, Object-Oriented Strategies and Intelligent Software Strategies. The former reports on commercial developments in object-oriented technologies while the latter reports on the commercial developments in the field of AI and expert systems, neural networks and intelligent CASE. He is also the co-author of Expert Systems: Artificial Intelligence in Business and two other books on expert systems. His latest book, ObjectCraft, has just been published by Addison-Wesley. Mr. Harmon manages his own consulting firm and consults with major corporations about their software systems strategies. He is also a popular speaker on the commercial impact of expert systems on business. He has given talks at conferences in Japan, France and the U.S.

Se June Hong has been the past general chair and program chair for this conference. He has led knowledge based systems effort at the IBM T. J. Watson Research Center for the past nine years. He is an lEEE fellow and has made contributions to the fault tolerant computing and design automation areas. He also serves as the U.S. delegate to IFIP TC-12 on AI.

Tom Laffey is co-founder and Vice President of Engineering for Talarian Corporation in Mountain View, California, where he directs a group in applying knowledge-based problem solving techniques to complex real-time systems. Before founding Talarian, he was a research scientist at Lockheed's Artificial Intelligence Center in Palo Alto, California. While there, Mr. Laffey was the project leader for Lockheed s L*STAR real-time knowledge-based architecture which has been applied to several real-time monitoring and control applications, including NASA s Hubble Space Telescope. Prior to coming to Lockheed, Mr. Laffey worked at the General Motors Research Laboratories from 1977 to 1980. Mr. Laffey has authored more than 30 articles in the area of knowledge-based systems. He received a M.S. in Computer Science from Virginia Polytechnic Institute and State University in 1982.

Aaron Marcus is an internationally recognized authority on graphic design for computer graphics. He and his staff have designed and evaluated knowledge visualization, user interfaces, and electronic publishing/presentations for many organizations as well as government clients. Mr. Marcus has written many articles on graphic design for computer graphics for technical and professional journals. He co-authored The Computer Image (1982), for which he wrote the essay Color: A Tool for Computer Graphics Communication, co-authored Human Factors and Typography for More Readable Programs (1990), and authored Graphic Design for Electronic Documents and User Interfaces (1991), all published by Addison-Wesley. Mr. Marcus received a B.A. in Physics from Princeton University (1965) and a B.F.A. and M.F.A. in Graphic Design from Yale University Art School (1968) . He has taught computer graphics since 1970 and founded AM+A in 1982.

Steven Minton is a member of the artificial intelligence research branch group at NASA Ames Research Center (employed by Sterling Federal Systems, a NASA contractor) . His primary research interests are in planning, scheduling, and learning. Currently, Dr. Minton is working on the application of learning techniques to real-world planning and scheduling problems at NASA. Prior to coming to NASA in September 1988, Dr. Minton was a post-doctoral research associate at Carnegie-Mellon University. He earned his Ph.D. in computer science from Carnegie-Mellon University in 1988. Dr. Minton was the primary architect of the PRODIGY system, an integrated planning and learning system. More recently, he has analyzed and extended a constraint-based scheduling algorithm used for scheduling the Hubble Space Telescope.

Bob O Keefe is an Associate Professor in the Department of Decision Sciences and Engineering Systems at Rensselaer Polytechnic Institute, and has a Ph.D. from the University of Southampton, England. Presently leading research into the validation and evaluation of knowledge based systems and their impact on organizations, he is also an advisor to the Food and Agriculture Organization of the United Nations on knowledge based system validation, and does work on the application of simulation and knowledge based methods in manufacturing. He is an associate editor of the International Journal of Intelligent Systems in Accounting, Finance and Management and the ACM Transactions on Modeling and Computer Simulation. His publications include papers in lEEE Expert, Expert Systems With Applications, the European Journal of Operational Research and the Journal of the Operational Research Society.

Dan 0 Leary is an Associate Professor in the School of Business at the University of Southern California, and has a Ph.D. from Case Western Reserve University. His current research interests include the verification, validation, assessment and security of knowledge based systems, and he is the editor of the International Journal of Intelligent Systems in Accounting, Finance and Management and an editorial board member of Expert Systems With Applications. He is the chair for AAAI sig AI in Business. His publications include papers in IEEE Expert, the European Journal of Operational Research, the International Journal of Man-Machine Systems and the International Journal of Expert Systems.

David S. Prerau is Principal Member of the Technical Staff at GTE Laboratories, Waltham, MA. He has led the successful development of several major industrial expert system. Dr. Prerau is the author of the well-regarded book Developing and Managing Expert Systems : Proven Techniques for Business and Industry. He has authored more than 70 published papers, reports, and book chapters including the most requested and reprinted article in the history of AI Magazine. He has presented tutorials to numerous corporate, professional and university audiences and has been an invited speaker nationally and internationally. Among his awards are the Outstanding Technical Paper Award from AFIPS and the Warner Technical Achievement Award from GTE, the corporation s highest technical award. Dr. Prerau received his Ph.D. in computer science from MIT.

Dr. Norman M. Sadeh is a member of both the Production Control Laboratory and the Enterprise Integration Laboratory at the Center for Integrated Manufacturing Decision Systems. He is currently the leader of the CORTES project, a project in decentralized production scheduling and control sponsored by DARPA and Digital Equipment Corporation. He is also the developer of the MICRO-BOSS factory scheduling system, a system that has yielded important reductions in scheduling costs (both tardiness and inventory costs) over more traditional bottleneck-centered approaches to factory scheduling. Dr. Sadeh s research interests include the applications of Artificial Intelligence and Operations Research techniques to both engineering and management problems. He received his Ph.D. in computer science from Carnegie Mellon University.

Carli Scott has 16 years of experience in the field of artificial intelligence and expert-systems technology. Her research and commercial experience spans a wide range of activities including: Building expert-system applications; Designing, implementing, documenting, testing, maintaining, and extending expert-system shells She worked for Stanford s Heuristic Programming project from 1974 to 1982 on the MYCIN and ONCOCCIN projects as a knowledge engineer and lead programmer, respectively. In 1981 Ms. Scott became a founder of Teknowledge Inc., and in 1982 she went to work for Teknowledge on a full-time basis as a Senior Knowledge Engineer where she worked on both application and product development. In 1990, she became a founder of Expert Support Inc.

Evangelos Simoudis is a research scientist in the Artificial Intelligence Center of Lockheed Corporation where he is directing research on case-based reasoning and inductive learning in manufactured settings. Dr. Simoudis received a Ph.D. in Computer Science from in 1991. His Ph.D. research resulted in the development of CASCADE and COAST, two case-based diagnostic systems that can be used in help desk settings where they reason from past cases captured by help desk engineers. Prior to joining Lockheed, Dr. Simoudis worked for six years as a Principal Engineer at Digital Equipment Corporation s Artificial Intelligence Center, where he conducted research on knowledge-based systems, explanation-based learning, and Distributed Artificial Intelligence.

Stephen F. Smith is a Research Scientist in the Robotics Institute at Carnegie Mellon University, and Director of the Production Control Laboratory within the Robotic Institute s Center for Integrated Manufacturing Decision Systems. He received his Ph.D. degree in Computer Science from the University of Pittsburgh in 1980. He has directed and participated in the development of several innovative knowledge-based scheduling systems, including the ISIS and OPIS factory scheduling systems and the HSTS Hubble Space Telescope observation scheduler, and has consulted on several other scheduling system development projects with various manufacturing and research organizations. Dr. Smith has published more than 35 articles in the area of knowledge-based planning, scheduling and control.

CAIA 92 WORKSHOP PROGRAM — Tuesday, March 3 9:00 am - 5:30 pm

Two by-invitation workshops will be held in conjunction with the conference. For general information contact Donald McKay, Unisys Defense Systems, Inc., P.O. Box 517, Paoli, PA 19301. ([email protected]) . 215-648-2256. WORKSHOP 1: Applying AI to Software Problems: Assessing Promises and Pitfalls

ORGANIZERS: Peter G. Selfridge, AT&T Bell Laboratories; Loren Terveen, AT&T Bel 1 Laboratories; William Mark, Lockheed Palo Alto Research Center; Richard M. Keller, NASA Ames Research Center

Problems in developing, maintaining, extending and understanding large software systems are becoming increasingly critical. This workshop seeks to bring together researchers in the areas of AI and Software Engineering to examine the promise and potential of applying AI to the real problems of software today, with three guiding questions. First, what are the real software problems and how can they be categorized? Second, what is the current state of AI and related Soft ware Engineering research with regard to maturity, scaling and related practical issues? Finally, are current research directions adequate to fulfill the promise of improved software productivity through AI? Of particular interest are lessons learned from success stories and constructive analysis of failures.

This workshop will be organized around a very small set of presentations. Each presentation will be followed by a discussant who will summarize the relevant issues and lead an open discussion. The goal is to foster intensive discussion on the problems of AI and Software among the participants.

Participants are invited from the AI and Software Engineering research communities, as well as people with experience in software development problems. Interested people should submit a position paper and a cover letter by January 1, 1992 indicating interest in being considered for making a presentation to: Peter G. Selfridge, AT&T Bell Laboratories, Room 28-425, Murray Hill, NJ 07974. Parti cipants will be notified of acceptance by February 1, 1992.

WORKSHOP 2: Artificial Intelligence for Customer Service and Support

ORGANIZERS: Anil Rewari, Digital Equipment Corp. (chair); Jeff Pepper, ServiceW are Inc.; Mike Register, Digital Equipment Corp.; Rajiv Enand, Carnegie Group, Inc.

This full-day workshop will gather researchers and developers who are explorin g or applying AI technology to the areas of customer service and support in industry as well as government. The areas of customer service and support are poised to be the leading areas for revenue growth to companies in the 90s. A 1991 market study shows that 23 percent of high technology service organizations have already begun building AI based systems and 67 percent expect to be using AI within the next three years.

The focus of this workshop will be on the use of AI techniques (including rule-based reasoning, case-based reasoning, model-based reasoning, neural nets, fuzzy logic, natural language processing and distributed AI) in developing intelligent tools and applications for customer service and support. These include, but are not limited to, knowledge-based troubleshooting systems (including diagnostic systems, advisory systems for service and support, voice response expert systems, knowledge acquisition tools and shells to build such systems) , intelligent information management systems (including text retrieval systems, call classification systems, case retrieval systems, interactive electronic technical manuals) , force planning and dispatch systems, maintenance planning applications, help desk applications and systems that analyze or use field back data.

Developers and researchers exploring and using AI techniques in the above areas of customer service and support are invited to participate in this workshop. Emphasis should be given to work that is ready for short term technology transfer rather than long term research issues. Prospective participants should send four (4) copies of an abstract that is not more than eight (8) pages long (not including title page and bibliography) double-spaced and in 12-point font with standard margins. In order to keep the workshop informal and to encourage active participation by all, the size of the workshop will be limited to 30 people. A subset of the participants will be selected as speakers. The deadline for submissions is January 10, 1992. Notification of invitation or rejection will be m ailed by January 24. Send four copies of each submission to Anil Rewari, Digita 1 Equipment Corp., 11l Locke Drive, LMO2-1/El2, Marlboro, MA 01752; Phone: 508/ 480-5568; Fax: 508/480-5822; Email: [email protected].

CAIA 92 HOTEL AND TRAVEL INFORMATION:

Doubletree Hotel at Fisherman's Wharf Two Portola Plaza Monterey, California 93940 408/649-4511

Reservations must be made under the group name of lEEE Computer Society CAIA 92 in order to receive the group rate. The group rate is $119 single or double occupancy, plus ten percent tax. A government rate of $73 single and $93 double will be extended if government identification is presented at check-in. Group rates will apply three days before and three days after the actual meeting dates. The reservation cut-off date is Saturday, February 1, 1992 at 5:00 p.m. A credit card or deposit of one night s room and tax is required to confirm a reservation. Deposits will be refunded by the hotel if notice of cancellation is received prior to 6:00 p.m. on the arrival date. Be sure to request either a confirmation number or written confirmation of any reservation or cancellation. Check-in time is 3:00 p.m. Guests may check in earlier subject to room availability. Check-out time is 12:00 p.m.

An arrangement has been made with United Airlines to provide group travel discounts of 5% off of the lowest available discounted fare or 45% off of the nonrestricted coach fare to either Monterey or San Francisco. The number to call for reservations is 800/521-4041. Refer to group I.D. #512 KP .

CAIA 92 CONFERENCE REGISTRATION

Return registration form to: CAIA 92 Registration, lEEE Computer Society, 1730 Massachusetts Avenue, N.W., Washington, DC 20036-1903, (202) 371-1013, FAX Number (202) 728-0884. Please Print or Type

Name Last/Family First M.I. Badge Name

Company Address /Mailstop City/State/Zip/Country Daytime Phone Number FAX Number lEEE/CS Membership Number

TUTORIALS March 2 and 3, 1992

#1 CASE, Knowledge-Based Systems and Object-Oriented Programming #2 Planning: Techniques and Applications #3 Case-Based Reasoning #4 Project Management and Development of Expert Systems in Business and Industry #5 Knowledge Acquisition Techniques #6 Practice Cannot Be Reduced to Theory - An Introduction to Social Science Perspectives on Software Design # 7 Developing Classification Rules (Trees) from Examples # 8 Knowledge-Based Production Management # 9 Real-Time Knowledge-Based Systems #10 Verification and Validation of Knowledge-Based Systems #11 Behavior-Based Reactive Robotic Systems #12 Graphical User Interface Design

TUTORIAL REGISTRATION (Price is per Tutorial)

Advance < 2/7/92 Late/On Site >2/7/92

Member : $190 Member: $230 Nonmember: $240 Nonmember: $285

CONFERENCE REGISTRATION

Advance < 2/7/92 Late/On Site > 2/7/92 Member: $280 Member: $340 Nonmember : $ 350 Nonmember: $425 Student: $105 Student: $115

Total Enclosed: $

Payment must be enclosed. Please make all checks payable to lEEE Computer Society. All payments must be in U.S. dollars, drawn on U.S. banks.

Method of payment accepted:

PersonalCheck Company Check Traveler's Check VISA MasterCard AmericanExpress Purchase Order (must accompany registration)

Cardholder Name Signature Card Number Expiration Date

Written requests for refunds must be received in the lEEE Computer Society office no later than February 7, 1992. Refunds are subject toa $50 processing fee. All no-show registrations will be billed in full. Students are required to show current picture ID cards at the time of registration. Registrations after February 14, 1992 will be accepted on-site only.

Nonmembers: Become a member today and register member prices! Call 202/371-1013 for details.