[Amdnl] CFP: Autonomous Agent Learning

[Gary Parker]

Autonomous Agent Learning
Special Session for the 2011 IEEE Congress on Evolutionary Computation 
(CEC 2011)
http://www.cec2011.org/

An autonomous agent has a self-sufficient control program that allows
it to operate in its environment. A pre-programmed controller,
although it may yield the desired results, is laborious to develop and
does not provide a means for adaptations necessitated by changes in
agent capability or the environment. Learning control through some
form of evolutionary computation can save development time and allow
the same control system to adapt to different environments, and in
some cases, adapt to changes in the agent’s capabilities and
environment.

Two areas where autonomous agents are of significant importance are in
robotics and interactive video games. Autonomous robots are used for
tasks that are too dangerous or impractical for humans to perform.
These include exploration in dangerous or remote locations,
intelligence gathering, and office or domestic assistance. Agents in
interactive video games are used to provide additional players (team
members or opponents) to interact with human players. Many of these
games are used purely for entertainment, but include games used for
training. In either case, having competitive, human like autonomous
agents in a game increases the realism and entertainment value. In
both autonomous robots and interactive video game agents, learning
systems can be significant aspects of their development and
enhancement.

This special session aims to discuss autonomous agent learning systems
with the primary emphasis on the use of evolutionary computation or
reinforcement learning to learn control programs for autonomous robots
and interactive video game agents. Topics for contributions include,
but may not be limited to the use of reinforcement learning or the
varying forms of evolutionary computation to learn controllers for
robots, simulations of robots, interactive video game agents, teams of
agents, and predator/prey agents.

Special session organizers:
Gary Parker, Connecticut College, USA (parker at conncoll.edu)
Hisashi Handa, Okayama University, Japan (handa at sdc.it.okayama-u.ac.jp)
Lee Graham, Connecticut College, USA (lgraham3 at conncoll.edu)

Submissions are still being accepted: http://www.cec2011.org/submission.htm
Select S19 as the main research topic.
Email any of the organizers if you have any questions.