This repository presents the BSI-PT (Bayesian Strategy Inference plus Policy Tracking) framework introduced in the paper Opponent Exploitation Based on Bayesian Strategy Inference and Policy Tracking.
BSI-PT is a Bayesian algorithm that can infer an opponent's policy in a multi-agent competitive environment. BSI-PT combines the advantages of inter-episode strategy inference and intra-episode policy tracking. Experiments have showed that BSI-PT is more accurate than other BPR variants at predicting the opponent's policy and winning against opponents with a variety of policy selection strategies.
- Kuei-Tso Lee
- Yen-Yun Huang
- Je-Ruei Yang
- Sheng-Jyh Wang
Please site our paper if you find this repository useful.
@article{lee2023opponent,
title={Opponent Exploitation Based on Bayesian Strategy Inference and Policy Tracking},
author={Lee, Kuei-Tso and Huang, Yen-Yun and Yang, Je-Ruei and Wang, Sheng-Jyh},
journal={IEEE Transactions on Games},
year={2023},
publisher={IEEE}
}Check the documentation to see how to run the experiments.