Abstract for AISB 2000: How to Design a Functioning Mind
AUTHOR: Doug Riecken,
IBM Watson Research Center & Rutgers University
TITLE We Must RE-MEMBER to RE-FORMULATE: The M System
ABSTRACT
In this paper I examine the following 2 problems. First, the
inter-working between different reasoning components and their
supporting resources to perform common sense reasoning and second, the
value of specific instinctive functions, "emotions", that serve as
"biasing resources" that in turn encode "positive and negative"
associations to encoded experiences, information, and knowledge.
In my earlier work in multi-strategy reasoning based on Minsky's Society
of Mind theories, I studied and prototyped an architecture that
demonstrated the development of multi-modal reasoning (e.g., structural,
functional, spatial, temporal, and causal). In this work, the system's
ability to perform causal reasoning was dependent upon its development
and maturity of other reasoning skills and necessary knowledge before it
could begin to perform causal reasoning. This study has continued with
the focus on the development of concept formation along with
categorization and naming. Specifically, what dependencies possibly
exist for a system to develop the ability to "learn" about "verb
concepts". What are the necessary components that specialize in
learning about "verb concepts" versus "noun concepts". The focus here is
on the representation and reasoning about concepts, not about
"language".
In earlier work, I examined and prototyped architectural components that
enable "learning" of "noun concepts" and the knowledge that was encoded
to represent such concepts. An evolution of this work now examines the
required components and their resources that enable "learning" of "verb
concepts". Specifically, (1) what specialized components enable concept
formation of "verb concepts" and (2) what are the other necessary
resources that serve as "supporting/complementing" components that
specialize in concept formation of "noun concepts" and their ontology of
"noun concepts". An important question is what level of maturity in
knowledge of "noun concepts" enables the learning of "verb concepts";
none, some, or a great deal. Another important question focuses on the
types of processing and representations that enable learning of "verb
concepts". In this work, I have made extensive application the Minsky
Trans-frame. Early output from this work has included a prototype
system that performs recognition and classification of tasks in a
micro-world.
With regard to "emotions", my work continues to study the role of a
system's resources that enable instinctive functions that evolve into
more complex behavioral functions. In this work and prototyping, I
continue to examine how "emotions" bias goal formulation, learning, and
problem solving.
=======================================================================
SHORT CV
Dr. Doug Riecken
Principal Investigator/Manager
Current: IBM Watson Research Center and Rutgers University
Prior: AT&T Bell Laboratories Research and Rutgers University
Ph.D. and M.S. studies under Marvin Minsky
Most Outstanding Paper - "Goal Formulation with Emotional Constraints:
Musical Composition by Emotional Computation". In AAAI Proceedings
First Annual Conference on Innovative Applications of Artificial
Intelligence, Stanford University, AAAI/MIT Press: Cambridge,
Massachusetts 1989.
Most Outstanding Paper - "Applying Agents in the Design of Multimedia
User Interfaces". In Proceedings of the ICAST-94 10th International
Conference on Advanced Science and Technologies , Naperville, Illinois
1994.
Keynote Address - ACM Annual International Conference on Intelligent
User Interfaces (IUI97), Orlando, Florida, 1997.
Keynote Address - AIMI/IEEE Conference on KANSEI: The Technology of
Emotion, Genova, Italy, 1997.
Keynote Address - Annual Metro Human Factors Society Conference,
Morristown, New Jersey, 1993.
Invited Paper - Workshop on Emotions in Humans and Artifacts, sponsored
by the Austrian Research Institute for Artificial Intelligence and the
Austrian Federal Ministry for Science, Vienna, Austria, 1999.
Invited Paper - ACM Conference on Electronic Technologies for Creativity
and Cognition, Loughborough University of Technology, U.K., 1993.
Guest Editor - Special Issue of the Communications of the ACM on
Intelligent Agents , July, 1994.
Member of Advisory Board of the Communications of the ACM
Member of Editorial Board of the Knowledge Based Systems Journal,
Elsevier
Conference Co-Chair for ACM/AAAI Annual Conference on Intelligent User
Interfaces (IUI2000), 2000.