Offline Transition Modeling via Contrastive Energy Learning

The official implementation of Offline Transition Modeling via Contrastive Energy Learning.

Installation

Download and install the main code from Energy Transition Model. The implementation is based on OfflineRL-Kit.

git clone https://github.com/Ruifeng-Chen/Energy-Transition-Models.git
cd Energy-Transition-Models
pip install -e .

Usage

Train an Energy Transition Model:

python run_example/train_etm.py

Evaluate policies within the Energy Transition Model:

python run_example/mbope.py --load_etm_path "your_ETM_path"

Perform offline policy optimization using EMPO within the learned energy transition models:

python run_example/run_empo.py --load_etm_path "ETM_path_1" "ETM_path_2" "ETM_path_3" "ETM_path_4" "ETM_path_5"

Citation

If you use this implementation in your work, please cite us with the following:

@inproceedings{
Energy Transition Model,
title={Offline Transition Modeling via Contrastive Energy Learning},
author={Ruifeng Chen and Chengxing Jia and Zefang Huang and Tian-Shuo Liu and Xu-Hui Liu and Yang Yu},
booktitle={Forty-first International Conference on Machine Learning},
year={2024},
url={https://openreview.net/forum?id=dqpg8jdA2w}
}