In IEEE World Congress on Computational Intelligence, pages 709-714, Ancourage, Alaska, May 1998.
The integration of different learning and adaptation techniques in one architecture has in recent years contributed to a large number of new intelligent system designs. We aim at classifying state-of-the-art intelligent systems and identify four categories, based on the systems' overall architecture: 1) single component systems, 2) fusion-based systems, 3) hierarchical systems, and 4) hybrid systems. We then introduce a unifying paradigm, derived from concepts well known in the AI and agent community, as conceptual framework to better understand, modularize, compare and evaluate the individual approaches. We believe it is crucial for the design of intelligent systems to focus on the integration and interaction of different learning techniques in one model rather then merging them to create ever new techniques. Two original instantiations of this framework are presented and discussed. Their performance is evaluated for prefetching of bulk data over wireless media.
Readers who enjoyed the above work, may also like the following: