Patrick F. Riley's Publications

Sorted by DateClassified by Publication TypeClassified by Research Category

Coaching Advice and Adaptation

Patrick Riley and Manuela Veloso. Coaching Advice and Adaptation. In Daniel Polani, Andrea Bonarini, Brett Browning, and Kazuo Yoshida, editors, RoboCup-2003: The Sixth RoboCup Competitions and Conferences, Springer Verlag, Berlin, 2004. (to appear)
Publisher's Webpage© Springer-Verlag

Download

[PDF]218.8kB  

Abstract

Our research on coaching refers to one autonomous agent providing advice to another autonomous agent about how to act. In past work, we dealt with advice-receiving agents with fixed strategies, and we now consider agents which are learning. Further, we consider agents which have various limitations, with the hypothesis that if the coach adapts its advice to those limitations, more effective learning will result. In this work, we systematically explore the effect of various limitations upon the effectiveness of the coach's advice. We state the two learning problems faced by the coach and the coached agents, and empirically study these problems in a predator-prey environment. The coach has access to optimal policies for the environment, and advises the predator on which actions to take. We experiment with limitations on the predator agent's actions, the bandwidth between the coach and agent, and the memory size of the agent. We analyze the results which show that coaching can improve agent performance in the face of all these limitations.

BibTeX

@InCollection(LNAI03-predprey,
  Author =	 "Patrick Riley and Manuela Veloso",
  Title =	 {Coaching Advice and Adaptation},
  booktitle =	 "{R}obo{C}up-2003: The Sixth {R}obo{C}up Competitions
                  and Conferences",
  Editor =	 {Daniel Polani and Andrea Bonarini and Brett Browning and Kazuo Yoshida},
  Publisher =	 "Springer Verlag",
  address =	 "Berlin",
  year =	 2004,
  wwwnote =	 {<a
                  href="http://www.springer.de/comp/lncs/index.html">Publisher's Webpage</a>&copy Springer-Verlag},
  note =	 {(to appear)},
  abstract =	 { Our research on coaching refers to one autonomous
                  agent providing advice to another autonomous agent
                  about how to act. In past work, we dealt with
                  advice-receiving agents with fixed strategies, and
                  we now consider agents which are learning. Further,
                  we consider agents which have various limitations,
                  with the hypothesis that if the coach adapts its
                  advice to those limitations, more effective learning
                  will result. In this work, we systematically explore
                  the effect of various limitations upon the
                  effectiveness of the coach's advice. We state the
                  two learning problems faced by the coach and the
                  coached agents, and empirically study these problems
                  in a predator-prey environment. The coach has access
                  to optimal policies for the environment, and advises
                  the predator on which actions to take. We experiment
                  with limitations on the predator agent's actions,
                  the bandwidth between the coach and agent, and the
                  memory size of the agent. We analyze the results
                  which show that coaching can improve agent
                  performance in the face of all these limitations.},
  bib2html_pubtype ={Refereed Conference},
  bib2html_rescat ={Coaching},
  bib2html_funding = {NSF,CoABS,ActiveTemplates},
)

Generated by bib2html.pl (written by Patrick Riley ) on Thu Mar 31, 2005 16:21:00