About 'Autonomous Intelligent Systems: Agents and Data Mining'

Since early 1990, multi-agent systems (MAS) and data mining and knowledge discovery (KDD) have remained areas of high interest for the research and development in intelligent information technologies. The multi-agent system technology offered powerful metaphors for information system conceptualization, a range of new architectures, techniques, and technologies specifically focused on the design and implementation of large-scale distributed intelligent systems operating in open heterogeneous environment. This technology is now approaching its industrial maturity, thus helping to cope with design, implementation, and deployment of critical industrial applications in many domains.

KDD technology also provides intelligent information technology with powerful ideas, algorithms, and software means intended for answering the main question of many Artificial Intelligence applied problems, "Where does the knowledge come from?", thus actually making modern applications intelligent and adaptive.

The evident recent trend in both science and industry is to integrate these technologies in order to use advantages of them both. The existing experience with combined application of multi-agent technology to design architectures of distributed data mining and KDD systems, and the utilization of distributed data mining and KDD to support learning tasks in multi-agent research, confirms the fact that both technologies are capable of mutual enrichment and their joint use results in information systems with emergent properties.

For instance, MAS architectures are used for distributed data mining and knowledge discovery systems, particularly, for distributed learning in data fusion systems making decisions based on many distributed heterogeneous data and information sources. The use of multi-agent technology in very large-scale data mining and KDD tasks is another example of the same trend. In return, multi-agent systems utilize data mining and KDD as a source of innovative ideas to provide applications more intelligence and adaptability.

The emergent technology may be outlined by such key words as autonomous intelligence, distributed intelligence, learning intelligence, emergent intelligence, emergent intelligent behavior, etc. It is worth noting that a number of very specific research problems arise on the cross-boundary of multi-agent, data mining and KDD technologies. The examples are specification framework and representation structures for distributed intelligence, the new role of ontology, joint use of knowledge extracted from different heterogeneous sources, protocols for distributed data mining, etc. Rapidly developing intelligent agent-based ad-hoc networks technology in particular, operating on peer-to-peer basis, represents one more example of the topmost problems currently given increasing attention from the scientific and industrial communities. Ubiquitous computing and ubiquitous knowledge discovery forming a class of novel distributed intelligent technology is practically infeasible without integration of multi-agent and KDD technologies.

During the past years, interest in joint use of the multi-agent and KDD technologies has also been proven by several successful international conferences and workshops which were devoted to the problem in question, either fully or partially.

The success of the First International Workshop "Autonomous Intelligent Systems: Agents and Data Mining" (AIS-ADM-05) held in June 2005 in St. Petersburg, Russia (http://space.iias.spb.su/ais05, LNAI-3505) confirmed the world wide interest in the theoretical and applied aspects of the new technology integrating agents, data mining and knowledge discovery.

However, like any newly emergent technology, integration of multi agent and data mining and KDD within distributed intelligent systems brought forward a number of novel problems and challenges. The Second International Workshop "Autonomous Intelligent Systems: Agents and Data Mining" (AIS-ADM-07) that will be held from June 3-5, 2007, in St. Petersburg, Russia, is intended for discovery and analysis of the aforementioned problems and challenges, as well as for sharing of new ideas, scientific results, lessons learned, and the best practices within the new cross boundary area integrating multi-agent and data mining and KDD technologies.

The Workshop invites researchers from both academia and industry to jointly discuss various aspects of theory and applications, as well as new opportunities provided by integration and interaction of multi-agent and data mining and KDD.