UNIVERSITY OF MASSACHUSETTS Computer Science Department
CMPSCI 791S	Sophisticated Agents in the Information Economy	Spring 2000
	Victor Lesser and Shlomo Zilberstein

Seminar Information

Description: The Internet is evolving into a large, free-market information economy in which autonomous agents can locate and extract useful information and use it solve problems and make decisions. This seminar will cover recent research directions aimed at building effective agent and multi-agent architectures for this complex environment and where appropriate will discuss implemented systems. Topics include: modeling information services, directory services and matchmaking, information extraction, retrieval and filtering, query planning and information integration, goal-directed and utility-directed information gathering, sharing knowledge and ontologies, negotiations and auctions, cost models and pricing of information, and the dynamics of information economies.

Meetings: Tuesday & Thursday 2:30-3:45, LGRC A339

Prerequisites: Some mathematical sophistication and at least one graduate level AI course (or permission of instructor). 3 credits.

Credit: 3 units

Instructors:
Victor Lesser, lesser@cs.umass.edu, 545-1322
Shlomo Zilberstein, shlomo@cs.umass.edu, 545-4189

Requirements: This seminar examines current research efforts aimed at developing autonomous systems that can locate, extract, and exploit targeted information from the vast amount of on-line data. It also explores the implications of such systems on the evolving information economy. The selected bibliography is divided into seven categories: (1) Introduction, (2) Directory services and matchmaking, (3) Information extraction, retrieval and filtering, (4) Query planning and information integration, (5) Sharing knowledge and ontologies, (6) Negotiations and auctions, and (7) The dynamics of information economies. In each meeting, we will discuss a few (1-3) papers that will be distributed in advance. Seminar discussions will be led by the instructors. Each participant is required to make two presentations of selected papers. In addition, each participant will have to prepare a short writeup for each meeting about the assigned reading. Grades will be based on the writeups, presentations, and class participation.

The papers for April 25 are: Philip R. Cohen and Hector J. Levesque, Communicative Actions for Artificial Agents. Les Gasser, Social Conceptions of Knowledge and Action: DAI Foundations and Open Systems Semantics. Artificial Intelligence, 47:107-138, 1991. Copies of this paper are available outside room 328.