SPEAKER: VICTOR LESSER

Professor of Computer Science and Director of the Multi-Agent Systems Laboratory, University of Massachusetts, Amherst


Evolution of Hearsay-II (Blackboards) as an Architecture for Interpretation Problems

ABSTRACT:
The Hearsay-II speech understanding system was the first implementation of a multi-level blackboard system. I will discuss extensions to this basic blackboard architecture over the last twenty-five years that has made it a more appropriate framework for constructing sophisticated interpretation systems that operate in open environments. As part of this discussion, I will describe a recently developed auditory scene analysis system, IPUS, which exploits an integrated search at all levels of the blackboard. Finally, I will touch on work that formalizes the asynchronous, multi-level search process that is at the heart of the blackboard architecture and provides a rigorous justification for many of the search control heuristics developed over the years.

SPEAKER BIO:
Victor Lesser received his A.B. in mathematics from Cornell University in 1966, and his M.S. and Ph.D. in computer science from Stanford University in 1969 and 1972, respectively. He was a research computer scientist at Carnegie-Mellon University from 1972 to 1976 working under the direction of Professor Raj Reddy. He joined the Computer Science faculty of University of Massachusetts, Amherst, Massachusetts in 1977, and is currently a Professor of Computer Science and Director of the Multi-Agent Systems Laboratory. His major research focus is on the control and organization of complex AI systems. He is a founding fellow of the American Association of Artificial Intelligence, and is considered a leading researcher in the areas of blackboard systems (he was the system architect for the Hearsay-II speech understanding system ), distributed AI/multi-agent systems (he is considered one of the founders of the field and was General Chairman of 1st International Conference on Multi-Agent Systems), and real-time AI (he developed the concept of approximate processing and its use in design-to-time scheduling). He has also made contributions in the areas of multiprocessor architectures, microprogramming, distributed operating systems, diagnostics, plan recognition, computer supported collaborative work, auditory scene analysis, and has recently been working on constructing agents for information gathering on WWW.

Return to Inventing the Future Home Page