Software Agents
Papers from the AAAI 1994 Spring Symposium
Oren Etzioni, ProgramChair
Technical Report SS-94-03
130 pp., $30.00
ISBN 978-0-929280-59-2
[Add to Cart] [View Cart]
Software agents are sensor/effector systems that operate within real-world software environments such as operating systems, databases, or computer networks. Their sensors observe features of this external environment, and their effectors can both alter the state of the environment directly, and communicate with other agents. Software agents pursue goals such as acquiring information about the environment or modifying its state, either individually or in teams. In contrast to work on human-computer collaboration, our focus is on agents with a high degree of autonomy and flexibility.
Advances in computer, information, and telecommunications technology have made software agents both necessary and possible. However, the role of AI and AI researchers in these developments has yet to be determined. We believe that AI has the potential to contribute to these developments, on the one hand, and that software domains offer fascinating research challenges on the other. To capitalize on this opportunity we seek to bring together researchers in this new area, to develop a common vocabulary, and to identify the fundamental research issues that define it. Specifically, we would like to identify:
- Challenge problems for the community (e.g. E-mail filtering)
- Possibilities for software agent interaction (e.g. can Maes and Mitchell's meeting scheduling systems query Etzioni's UNIX softbot for information about potential meeting participants such as their phone numbers, current location, etc.?)
- Possible connections to computer scientists outside AI (in databases, operating systems, office automation, etc.)