Distributed artificial intelligence meets machine learning. Learning in multi-agent environments. ECAI '96 workshop LDAIS, Budapest, Hungary, August 13, 1996, ICMAS '96 workshop LIOME, Kyoto, Japan, December 10, 1996. Selected papers (Q1356014)

This book contains sixteen selected, revised and extended versions of papers originally given at two workshops whose common theme was the intersection of distributed artificial intelligence and machine learning methodologies, viz. The ECAI'96 workshop on Learning in Distributed Artificial Intelligence Systems, held in Budapest, Hungary, in August 1996, and the ICMAS'96 workshop on Learning, Interaction, and Organization in Multiagent Environments, held in Kyoto, Japan, in December 1996. The invited paper by Singh and Huhns provides on overview and a discussion of major challenges for machine learning in cooperative information systems: the extraction of semantics, coordination and collaboration, and abstractions and structure of cooperative information systems. The remaining sixteen technical papers can be grouped into three major categories: The first of these relates to learning, cooperation and competition. This part includes work whose main focus is on how multiple agents can learn collaborative and cooperative behavior (e.g., in order to optimally share resources or to maximize one's own profit). This work is much concerned with the development and adaptation of data flow and control patterns that improve the interaction among several agents (e.g., by applying reinforcement learning algorithms in a multi-agent setting). The second category subsumes the issue of learning about/from other agents and the world; application scenarios range from robots learning to play soccer to organizational learning, as exemplified by task allowcations problems. Here the main focus is on how learning conducted by an agent can be influenced (e.g., initiated, accelerated, re-directed) by other agents. This work (which shows several applications of the contact net protocol) is much concerned with the prediction of the behavior of other agents (including their preferences, strategies, intentions, goals), with the improvement and refinement of an agent's behavior by interacting with and observing other agents, and with the development of a common view (ontology, epistemological foundations) of the world. The third and final category covers learning, communication and understanding. Here the main focus is on how learning, on the one hand, and communication and mutual understanding, on the other hand, are related to each other. This work is much concerned with the requirements of the agents' ability to effectively exchange useful information and to develop a shared meaning of the information exchanged.