/H-KGA

Primary LanguagePythonMIT LicenseMIT

H-KGA

arch

This repository contains official implementation for our paper:

Generalization in Text-based Games via Hierarchical Reinforcement Learning

Yunqiu Xu, Meng Fang, Ling Chen, Yali Du and Chengqi Zhang

Dependencies

Our code depends heavily on xingdi-eric-yuan/GATA-public. Specifically, we use the version 413a262. The requirements could be found at requirements.txt.

How to train

  • Download the word embeddings:
wget "https://bit.ly/2U3Mde2"
  • Download the games:
    • L3/L7/L5/L9 correspond to S1/S2/S3/S4
    • L1/L2/L4/L8 correspond to US1/US2/US3/US4
wget https://aka.ms/twkg/rl.0.1.zip
python train_rl.py train_rl_config.yaml

Citation

@article{xu2021hierarchical,
  title={Generalization in Text-based Games via Hierarchical Reinforcement Learning},
  author={Xu, Yunqiu and Fang, Meng and Chen, Ling and Du, Yali and Zhang, Chengqi},
  journal={arXiv preprint arXiv:2109.09968},
  year={2021}
}

License

MIT License