Designing Agents for Context-Rich Textual Information Tasks
Much of information and knowledge are documented in free texts. Textual task capabilities and agencies are inevitably essential to successful information services. In this paper, we describe some empirical observations on information tasks in a context rich data domain and attempt to discuss its implications on agent system design. We have developed an agent system with two essential task capabilities - information retrieval and information extraction, which can be built upon for more value-added information services. We observed a number of textual information task characteristics, such as process-centric, independently decomposable operations, indispensable domain knowledge, and user driven task specification, that are influential to designing agent systems. We also propose a conceptual view on system design that considers three agent groups - task agent, knowledge agent, and operation agent. The characterization of agent roles helps determine primary functions needed and set apart stable intermediate forms in the system. Further analysis on relationship among components would reveal major types and patterns of interaction and how the agents should be designed to coordinate with each other.
Unable to display preview. Download preview PDF.
- 1.Bayardo, R. J. et.al. InfoSleuth: agent-based semantic integration of information in open and dynamic environments. Proceedings of the ACM SIGMOD Int’l Conf. on Management of Data, pp. 195–206, 1997.Google Scholar
- 3.Durfee, E.H., Kiskis, D.L., and Birmingham, W.P. The Agent Architecture of the University of Michigan Digital Library. In Readings in Agents, Huhns & Singh, (Eds.) pp. 98–110, 1998.Google Scholar
- 4.Etzioni, O. Moving Up the Information Food Chain. AI Magazine, vol. 18(2) pp. 11–18, 1997.Google Scholar
- 5.Genesereth, M. An Agent-Based Framework for Interoperability. In Readings in Agents, Huhns & Singh, (Eds.) pp. 317–346, 1998.Google Scholar
- 6.Grishman, R. Information Extraction: Techniques and Challenges. In Information Extraction-A Multidisciplinary Approach to art Emerging Information Technology. 10–27, 1997.Google Scholar
- 7.Hayes-Roth, F. Artificial Intelligence-What works and What Doesn’t? AI Magazine, vol. 18(2)pp. 99–113, 1997.Google Scholar
- 8.Jennings, N. and Wooldridge, M. Software Agents. IEE Review, pp 17–20, January 1997.Google Scholar
- 9.Liu, J., Soo, V. Chiang, C. et al. Gaz-Guide: Agent-Mediated Information Retrieval for Official Gazettes. In Intelligent Agents: Specification, Modeling, and Applications, Yuan & Yokoo (Eds.), (PRIMA 2001), LNAI 2132, Springer, 2001.Google Scholar
- 10.Maes, M. Agents that reduce work and information overload. Comm. of the ACM, vol. 37(7) pp. 31–40, July 1994.Google Scholar
- 11.Message Understanding Conference, http://www.itl.nist.gov/iaui/894.02/related_projects/muc/
- 12.Salton, G. and McGill, M. J. Introduction to Modern Information Retrieval. McGraw-Hill, Inc., 1983.Google Scholar
- 13.Simon, H. The Science of the Artificial. MIT Press, 1996.Google Scholar
- 14.Sycara, K., et. al. Distributed Intelligent Agents. IEEE Expert, vol. 11(6), 1996.Google Scholar