IJCAI-97 HOME

IJCAI-97 Invited Talks


Index

Let's Plan it Deductively!
Wolfgang Bibel

Creativity and Artificial Intelligence
Margaret A. Boden

Modeling Social Action for AI Agents
Cristiano Castelfranchi

Vehicles Capable of Dynamic Vision
Ernst D. Dickmanns

Remote-Brained Robots
Masayuki Inaba

Generating Multimedia Briefings:
Language Generation in a Coordinated Multimedia Environment to Convey Information Concisely

Kathleen R. McKeown

Inheritance Comes of Age:
Applying Nonmonotonic Techniques to Problems in Industry

Leora Morgenstern

Machine Learning Techniques to Make Computers Easier to Use
Hiroshi Motoda

The New Millennium Remote Agent:
To Boldly Go Where No AI System Has Gone Before

Nicola Muscettola, P. Pandurang Nayak, Barney Pell and Brian C. Williams

Reinforcement Learning: Lessons for AI
Richard S. Sutton

Numerica: A Modeling Language for Global Optimization
Pascal Van Hentenryck

The Origins of Syntax in Visually Grounded Robotic Agents
Luc Steels


Let's Plan it Deductively!

Wolfgang Bibel
Technical University Darmstadt, Germany

abstract

Logic and its deductive machinery is in high academic regard but is more or less irrelevant in industry. Even systems supporting an obviously logical task such as planning rarely resort to logic and deduction. Neither truly intelligent planning nor artificial intelligence in general will ever be achieved unless deductive mechanisms are given a more central role. State, results and issues of deduction and deductive planning, and why these two areas matter for AI, are thus the topics of this talk.

Creativity and Artificial Intelligence

Margaret A. Boden
University of Sussex, United Kingdom

abstract

Creativity is a fundamental feature of human intelligence, and a challenge for AI. AI techniques can be used to create new ideas in three ways: by producing novel combinations of familiar ideas; by exploring the potential of conceptual spaces; and by making transformations that enable the generation of previously impossible ideas. AI will have less difficulty in modelling the generation of new ideas than in automating their evaluation.

Modeling Social Action for AI Agents

Cristiano Castelfranchi
National Research Council and University of Siena, Italy

abstract

Agent-Based Computing needs social intelligence. Some basic issues for understanding and designing social agents will be addressed: the difference between social and collective, as well as between interaction and communication; the trade-off between autonomy and compliance with requests, norms and constraints; the relationships between deliberate cooperation and emergent intelligence and functions; and the foundational role of goal-delegation and goal-adoption for defining different levels of agency and cooperation.

Vehicles Capable of Dynamic Vision

Ernst D. Dickmanns
University of the German Army at Munich, Germany

abstract

A survey will be given on two decades of developments in the field, encompassing an increase in computing power by four orders of magnitude. The '4-D approach' integrating expectation-based methods from systems dynamics and control engineering with methods from AI has allowed the creation of vehicles with unprecedented capabilities in the technical realm: autonomous road vehicle guidance in public traffic on freeways at speeds beyond 130 km/h, on-board-autonomous landing approaches of aircraft, and landmark navigation for AGV's, for road vehicles including turn-offs onto cross-roads, and for helicopters in low-level flight (real-time, hardware-in-the-loop simulations in the latter case).

Remote-Brained Robots

Masayuki Inaba
The University of Tokyo, Japan

abstract

AI and Robotics once shared a dream. Technical advances have led to an age where this dream seems within reach; but it also seems that recent AI and robotics may fail to inspire the next generation of researchers. Remote-Brained Robots may provide a way to reinvent and revitalize the AI dream; this talk introduces the approach.

Generating Multimedia Briefings:
Language Generation in a Coordinated Multimedia Environment to Convey Information Concisely

Kathleen R. McKeown
Columbia University, USA

abstract

Communication can be more effective when several media (such as text, speech, or graphics) are integrated and coordinated to present information. This changes the nature of media specific generation (e.g., language generation) which must take into account the multimedia context in which it occurs. In this talk, I will present work on coordinating and integrating speech, text, static and animated 3D graphics, and stored images, as part of several systems we have developed at Columbia University. A particular focus of our work has been on the generation of presentations that brief a user on information of interest.

Inheritance Comes of Age:
Applying Nonmonotonic Techniques to Problems in Industry

Leora Morgenstern
IBM T. J. Watson Research Center, USA

abstract

Inheritance has long been considered the stepchild of formal AI: a quick and convenient but unexciting way to model some of the more routine patterns of commonsense reasoning. However, a simple change in the inheritance paradigm - attaching formulas to the nodes in a network - allows us to use semantic networks to do general default reasoning. Moreover, we can retain many of the positive features of inheritance networks to solve difficult default reasoning problems in a straightforward manner. This talk presents the new paradigm and describes how it has been used for real applications in the medical insurance and life insurance industries.

Machine Learning Techniques to Make Computers Easier to Use

Hiroshi Motoda
Osaka University, Japan

abstract

Computers are still not easy to use. Difficulties come from their ignorance about the user. Identifying user-dependent information that can be automatically collected helps build a part of user model by which to predict what the user wants to do next and to do relevant preprocessing. How this is made possible by using machine learning techniques is the topic of this talk. This is joint work with Kenichi Yoshida of the Advanced Research Laboratory, Hitachi, Japan.

The New Millennium Remote Agent:
To Boldly Go Where No AI System Has Gone Before

Nicola Muscettola
Recom Technologies, NASA Ames Research Center, US

P. Pandurang Nayak
Recom Technologies, NASA Ames Research Center, US

Barney Pell
Caelum Research Corporation, NASA Ames Research Center, US

Brian C. Williams
Recom Technologies, NASA Ames Research Center, US

abstract

The New Millennium Remote Agent (NMRA) is an autonomous spacecraft control system being developed jointly by NASA Ames and JPL. It integrates constraint-based planning and scheduling, robust multi-threaded execution, model-based diagnosis and reconfiguration, and real-time monitoring and control. NMRA will control Deep Space One (DS-1), the first flight of NASA's New Millennium Program (NMP), which will launch in 1998. As the first AI system to autonomously control an actual spacecraft, NMRA will enable the establishment of a "virtual presence" in space through an armada of intelligent space probes that autonomously explore the nooks and crannies of the solar system.

Reinforcement Learning: Lessons for AI

Richard S. Sutton
University of Massachusetts at Amherst, USA

abstract

The field of reinforcement learning has recently produced world-class applications and, as we survey in this talk, scientific insights that may be relevant to all of AI. In my view, the main things that we have learned from reinforcement learning are 1) the power of learning from experience as opposed to labeled training examples, 2) the central role of modifiable evaluation functions in organizing sequential behavior, and 3) that learning and planning could be radically similar.

Numerica: A Modeling Language for Global Optimization

Pascal Van Hentenryck
Brown University, USA

abstract

Numerica is a modeling language for stating and solving nonlinear problems over the reals. From a syntactic standpoint, Numerica makes it possible to state nonlinear problems as in textbooks or scientific papers. From a semantic standpoint, Numerica is guaranteed to locate all isolated solutions to nonlinear constraint systems and all global optima in optimization problems. From an implementation standpoint, Numerica combines methods from numerical analysis and constraint satisfaction.

The Origins of Syntax in Visually Grounded Robotic Agents

Luc Steels
VUB AI Laboratory, Belgium and Sony Computer Science Laboratory, France

abstract

Experiments are reported in which a group of software and/or robotic agents are able to develop a shared set of conventions with the multi-layered structure and complexity of natural languages. The languages are grounded in the environmental and bodily experiences as perceived by the agents. It is further shown how there can be a co-evolution of language and meaning and hence a progressive build up of cognitive competence.
higuchi@etl.go.jp
Last modified: Tue Jun 10 22:00:16 JST 1997