The instructors are Cindy Mason, PhD, School of Medicine, Stanford
University, Henry Lieberman, PhD, MIT Media Laboratory, and
Ted Selkirk, PhD, IBM Almaden Research Center.
Agent technology is one of the fastest growing areas of computer and
information science. Intelligent software agents are essential to manage
the growing complexity and volume of information in our environment.
Recently, agents have been created that can act on our behalf to sort
electronic mail, navigate and retrieve information on the Internet,
for resources, and schedule meetings. In the future, agents may navigate
and drive automobiles, or perform shopping functions according to budget
constraints and personal preferences.
This course explores the design, implementation, and use of two types of
computational agents: interface agents and collaborative agents. Interface
agents support the learning and use of computer tools, such as operating
systems. An interface agent can observe the actions taken by the user in
the interface, learn new capabilities dynamically, suggest courses of action
to the user, provide context-sensitive help, adapt the interface
to the user's personalized requirements, or automate tasks that would
otherwise require tedious sequences of manual operations. The design
of interface agents draws on representation, learning, and reasoning
techniques used in other branches of artificial intelligence, but carefully
integrates principles of good human-computer interaction.
Collaborative agents interact and cooperate with other agents to perform
tasks on behalf of a user. The design of collaborative agents involves
problem solving, communication, and coordination strategies for agents
to maintain autonomy, yet benefit from the network as a whole. Agent
collaborations may involve heterogeneous or homogeneous groups
of agents, and agents with similar or differing goals, languages, and
knowledge representation facilities. Collaborative agent technology
draws on principles of artificial intelligence, sociology, organizational
theory, animal behavior, economics, and distributed systems.
This course describes the utility of current and experimental software
agents as assistants and advisors, as well as the technologies involved
in their design. In addition, the course examines the social implications
of software agents. Agents are discussed in the contexts of electronic
mail, browsing the World Wide Web, digital libraries, graphical editors,
and tutoring systems. Current agent software and preview future agent
technology are also demonstrated.
The course fee is $1295, which includes course materials.
For additional information and a complete course description, please
contact Marcus Hennessy at:
(310) 206-2815 fax