View Publication

A Hierarchy of Prescriptive Goals for Multiagent Learning

Martin Zinkevich
A Greenwald
Michael L. Littman, AT&T Labs-Research, Florham Park, New Jersey

A great deal of theoretical effort in multiagent learning involves either embracing or avoiding the inherent symmetry between the problem and the solution. Regret minimization is an approach to the prescriptive, non-cooperative goal that explicitly breaks this symmetry, but, since it makes no assumptions about the adversary, it achieves only limited guarantees. In this paper, we consider a hierarchy of goals that begins with the basics of regret minimization and moves towards the utility guarantees achievable by agents that could also guarantee converging to a game-theoretic equilibrium.

Citation

M. Zinkevich, A. Greenwald, M. Littman. "A Hierarchy of Prescriptive Goals for Multiagent Learning". Artificial Intelligence (AIJ), 171, pp 440-447, May 2007.

Keywords:	machine learning
Category:	In Journal

BibTeX

@article{Zinkevich+al:AIJ07,
  author = {Martin Zinkevich and A Greenwald and Michael L. Littman},
  title = {A Hierarchy of Prescriptive Goals for Multiagent Learning},
  Volume = "171",
  Pages = {440-447},
  journal = {Artificial Intelligence (AIJ)},
  year = 2007,
}

Last Updated: August 20, 2007
Submitted by Russ Greiner

Not Logged In

PapersDB

A Hierarchy of Prescriptive Goals for Multiagent Learning

Citation

BibTeX