Skip to content

yinxusen/deepword

Repository files navigation

██████╗ ███████╗███████╗██████╗ ██╗    ██╗ ██████╗ ██████╗ ██████╗
██╔══██╗██╔════╝██╔════╝██╔══██╗██║    ██║██╔═══██╗██╔══██╗██╔══██╗
██║  ██║█████╗  █████╗  ██████╔╝██║ █╗ ██║██║   ██║██████╔╝██║  ██║
██║  ██║██╔══╝  ██╔══╝  ██╔═══╝ ██║███╗██║██║   ██║██╔══██╗██║  ██║
██████╔╝███████╗███████╗██║     ╚███╔███╔╝╚██████╔╝██║  ██║██████╔╝
╚═════╝ ╚══════╝╚══════╝╚═╝      ╚══╝╚══╝  ╚═════╝ ╚═╝  ╚═╝╚═════╝ 

DeepWord

Create auto-agent to play text-based games

Our testbed is TextWorld. Our BaseAgent is built upon TextWorld. However, our agent core is independent with specific game frameworks or games.

The BaseAgent works with specific games/frameworks, fetching important information from games, and gives actions as feedback.

The BaseCore works independently, requiring only trajectory and actions, etc.

Read the tutorial.md for users.

Check the API doc for developers.

Cite our papers

@article{DBLP:journals/corr/abs-1905-02265,
  author    = {Xusen Yin and
               Jonathan May},
  title     = {Comprehensible Context-driven Text Game Playing},
  journal   = {CoRR},
  volume    = {abs/1905.02265},
  year      = {2019},
  url       = {http://arxiv.org/abs/1905.02265},
  archivePrefix = {arXiv},
  eprint    = {1905.02265},
  timestamp = {Mon, 27 May 2019 13:15:00 +0200},
  biburl    = {https://dblp.org/rec/bib/journals/corr/abs-1905-02265},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

@misc{yin2019learn,
    title={Learn How to Cook a New Recipe in a New House: Using Map Familiarization, Curriculum Learning, and Bandit Feedback to Learn Families of Text-Based Adventure Games},
    author={Xusen Yin and Jonathan May},
    year={2019},
    eprint={1908.04777},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

@misc{yin2020zeroshot,
    title={Zero-Shot Learning of Text Adventure Games with Sentence-Level Semantics},
    author={Xusen Yin and Jonathan May},
    year={2020},
    eprint={2004.02986},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

@misc{yin2020learning,
      title={Learning to Generalize for Sequential Decision Making}, 
      author={Xusen Yin and Ralph Weischedel and Jonathan May},
      year={2020},
      eprint={2010.02229},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

About

reinforcement learning for text-based games

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages