Abalearn: A Risk-Sensitive Approach to Self-Play Learning in Abalone
Abstract
This paper presents Abalearn, a self-teaching Abalone program capable of automatically reaching an intermediate level of play without needing expert-labeled training examples, deep searches or exposure to competent play. Our approach is based on a reinforcement learning algorithm that is risk-seeking, since defensive players in Abalone tend to never end a game. We show that it is the risk-sensitivity that allows a successful self-play training. We also propose a set of features that seem relevant for achieving a good level of play. We evaluate our approach using a fixed heuristic opponent as a benchmark, pitting our agents against human players online and comparing samples of our agents at different times of training.
Cite
Text
Campos and Langlois. "Abalearn: A Risk-Sensitive Approach to Self-Play Learning in Abalone." European Conference on Machine Learning, 2003. doi:10.1007/978-3-540-39857-8_6Markdown
[Campos and Langlois. "Abalearn: A Risk-Sensitive Approach to Self-Play Learning in Abalone." European Conference on Machine Learning, 2003.](https://mlanthology.org/ecmlpkdd/2003/campos2003ecml-abalearn/) doi:10.1007/978-3-540-39857-8_6BibTeX
@inproceedings{campos2003ecml-abalearn,
title = {{Abalearn: A Risk-Sensitive Approach to Self-Play Learning in Abalone}},
author = {Campos, Pedro F. and Langlois, Thibault},
booktitle = {European Conference on Machine Learning},
year = {2003},
pages = {35-46},
doi = {10.1007/978-3-540-39857-8_6},
url = {https://mlanthology.org/ecmlpkdd/2003/campos2003ecml-abalearn/}
}