Proceedings of the workshop "Adaptive Systems and User Modeling on the World Wide Web",
Sixth International Conference on User Modeling, Chia Laguna, Sardinia, 2-5 June 1997

DIAMS: Distributed Intelligent Agents for Information Management and Sharing

Nathalie Mathe and James R. Chen
Recom Technologies, Inc.
Computational Sciences Division
NASA Ames Research Center
Moffett Field, CA 94035-1000
mathe@ptolemy.arc.nasa.gov

Abstract: The Internet and Intranet revolution has made a wealth of information resources available for direct and easy access on the user's desktop. However, finding appropriate information has become a significant problem for many users. Our research focuses on three areas which require significant technological advances: finding information relevant to users' needs, organizing information for facilitating access in various contexts, and collaborative information sharing. We propose to develop next generation distributed and collaborative information agents (DIAMS) which support dynamic and flexible organization of personal information repositories, distributed over the WWW. These in turn enable sharing between users at the knowledge-level, and automated discovery of new relevant information through collaborative information exchange between software agents. The DIAMS system will learn from users' preferences for organizing information, and provide dynamic views of information according to current users' needs. DIAMS will integrate knowledge-based, neural networks and genetic algorithms technologies.

Keywords: information access, information organizing, virtual folders, collaborative information sharing, user preferences, user interests, neural networks, knowledge-based systems

Introduction

The Internet and Intranet revolution has made a wealth of information resources available for direct and easy access on the user's desktop. However, finding appropriate information has become a significant problem for many users, as well as for many NASA workers or team of scientists. Our research focuses on three areas which require significant technological advances: finding information relevant to users' needs, organizing information for facilitating access in various contexts, and collaborative information sharing. Current WWW search engines allow users to locate information of interest to some extend, but often return vast amount of irrelevant information. More recent user profile and collaborative filtering technologies attempt to provide relevant information to users by learning from their previous queries or from other users' queries and feedback, but results are still of insufficient quality. Yet users need an easy way to access information relevant and adapted to their current task and interest at any time.

Moreover, once relevant information is found, pointers to it must be locally organized and stored in a manner that allows rapid and effective access for both individuals and workgroups. Current information organizing schemes are mostly static, hierarchical, and monolithic. They do not support fast changing rate of information, nor multiple uses of information under various contexts or by various users. There is also a critical need for tools supporting collaboration among distributed workgroups. Users need both to collectively gather and share information relevant to the group's task. Sharing a common repository of information is a first step, but doesn't scale up to large distributed dynamic groups. Collaborative tools themselves need to be distributed and dynamic, and support collaborative learning and discovery of information.

Collaborative Information Agents

In order to improve access, sharing and organization of information, we propose to develop next generation distributed and collaborative information agents (DIAMS) which support dynamic and flexible organization of personal information repositories, distributed over the WWW. These in turn enable sharing between users at the knowledge-level, and automated discovery of new relevant information through collaborative information exchange between software agents. The DIAMS system will learn from users' preferences for organizing information, and provide dynamic views of information according to current users' needs. DIAMS will integrate knowledge-based, neural networks and genetic algorithms technologies. It will also be able to interact with other tools of leading edge information technologies.

We already developed the current generation of adaptive indexing and retrieval agent (ARNIE), which enable users to categorize and share information of interest under various contexts, and is able to learn users' interests based on how they organize information, rather than what they search for [Mathe & Chen, 96]. ARNIE encodes information relevance and structure in a neural network dynamically configured with a genetic algorithm. This core technology was successfully integrated into several applications: Boeing Portable Maintenance Aid prototype, NASA/JSC Adaptive HyperMan Electronic Documentation system in use at JSC Mission Control [Rabinowitz, Mathe & Chen, 95], and WebTagger(TM) (an adaptive bookmarking service on the Web) [Keller, et al., 97]. WebTagger is a personal bookmarking service that provides both individuals and groups with a customizable means of organizing and accessing Web-based information resources. In addition, the service enables users to supply feedback on the utility of these resources relative to their information needs, and provides dynamically-updated ranking of resources based on incremental user feedback. Individuals may access the service from anywhere on the Internet.

The next generation collaborative agents will extend the current architecture with multi-attribute network objects, added semantic knowledge and symbolic processing capabilities. We propose to develop new information access methods which provide dynamically organized views of personal information repositories using knowledge-based and neural network representation and indexing. We will develop a knowledge exchange protocol for intelligent collaboration, and implement advanced techniques for sharing and gathering of information using knowledge-based, distributed automated agents. Finally, the new research design puts emphasis on a modular architecture with flexible interface protocols. The system can hence take advantage of existing tools with capabilities such as text search and automated categorization, and interact with established on-line dictionaries, thesauri, and knowledge bases.

References

[Keller, et al., 97]
Keller, R. M., Wolfe, S. R., Chen, J. R., Rabinowitz, J. L., and Mathe, N. (1997); A Bookmarking Service for Organizing and Sharing URLs; Proceedings of the Sixth International WWW Conference, Santa Clara, CA, Apr. 1997.
[Mathe & Chen, 96]
Mathe, N. and Chen, J. R. (1996); User-Centered Indexing for Adaptive Information Access; International Journal of User Modeling and User Adapted Interaction, Special Issue on Adaptive Hypertext and Hypermedia 6(2-3):225-261.
[Rabinowitz, Mathe & Chen, 95]
Rabinowitz, J., Mathe, N. & Chen, J. R. (1995); Adaptive HyperMan: A Customizable Hypertext System for Reference Manuals; Proceedings of the AAAI Fall Symposium on Artificial Intelligence Applications in Knowledge Navigation and Retrieval, Cambridge, MA, Nov. 10-12. [Postcript version- 738 Kb]