Convergence, Targeted Optimality, and Safety in Multiagent Learning

Abstract

This paper introduces a novel multiagent learning algorithm, Convergence with Model Learning and Safety (or CMLeS in short), which achieves convergence, targeted optimality against memory-bounded adversaries, and safety, in arbitrary repeated games. The most novel aspect of CMLeS is the manner in which it guarantees(in a PAC sense) targeted optimality against memory-bounded adversaries, via efficient exploration and exploitation. CMLeS is fully implemented and we present empirical results demonstrating its effectiveness.

Cite

Text

Chakraborty and Stone. "Convergence, Targeted Optimality, and Safety in Multiagent Learning." International Conference on Machine Learning, 2010. doi:10.1007/978-3-319-02606-0_4

Markdown

[Chakraborty and Stone. "Convergence, Targeted Optimality, and Safety in Multiagent Learning." International Conference on Machine Learning, 2010.](https://mlanthology.org/icml/2010/chakraborty2010icml-convergence/) doi:10.1007/978-3-319-02606-0_4

BibTeX

@inproceedings{chakraborty2010icml-convergence,
  title     = {{Convergence, Targeted Optimality, and Safety in Multiagent Learning}},
  author    = {Chakraborty, Doran and Stone, Peter},
  booktitle = {International Conference on Machine Learning},
  year      = {2010},
  pages     = {191-198},
  doi       = {10.1007/978-3-319-02606-0_4},
  url       = {https://mlanthology.org/icml/2010/chakraborty2010icml-convergence/}
}