siddharthnishtala / cost-sensitive-trees-rl

Learn decision tree policies in reinforcement learning.

Cost-Sensitive Trees for Interpretable Reinforcement Learning

This repository contains the source code accompanying the paper Cost-Sensitive Trees for Interpretable Reinforcement Learning.

Installation

The codebase uses slightly modified versions of scikit-learn and stable-baselines3 libraries. Building scikit-learn from source requires other dependencies. Please follow instructions from here to setup before installing these libraries.

Once this is setup, the libraries and all other dependencies can be installed using:

pip install -r requirements.txt

The code was written and tested on Python 3.7.9.

BibTeX Citation

If you use CS-VIPER or CS-MoET in a scientific publication, we would appreciate using the following citation:

@inproceedings{cost-sensitive-trees-rl,
    author = {Nishtala, Siddharth and Ravindran, Balaraman},
    title = {Cost-Sensitive Trees for Interpretable Reinforcement Learning},
    year = {2024},
    isbn = {9798400716348},
    publisher = {Association for Computing Machinery},
    address = {New York, NY, USA},
    url = {https://doi.org/10.1145/3632410.3632443},
    doi = {10.1145/3632410.3632443},
    booktitle = {Proceedings of the 7th Joint International Conference on Data Science \& Management of Data (11th ACM IKDD CODS and 29th COMAD)},
    pages = {91–99},
    numpages = {9},
    keywords = {interpretability, reinforcement learning, decision trees},
    location = {Bangalore, India},
    series = {CODS-COMAD '24}
}