Skip to content
/ LoFT Public

Code for EACL 2023 paper "LoFT: Enhancing Faithfulness and Diversity for Table-to-Text Generation via Logic Form Control"

License

Notifications You must be signed in to change notification settings

Yale-LILY/LoFT

Repository files navigation

LoFT

The data and code for EACL 2023 paper LoFT: Enhancing Faithfulness and Diversity for Table-to-Text Generation via Logic Form Control, which applies logical forms as fact checkers and content planners to improve faithfulness and diversity of logical table-to-text generation simultaneously.

Main Modules

The LoFT code is organized into the following three modules (in order of execution):

  • LoFT_data_processing: which prepares the training and inference data for LoFT.
  • LoFT_framework: which contains the implementation of training and inference process of LoFT.
  • LoFT_evaluation: which contains the implementation of evaluation metrics of LoFT.

We provide details of each module in README.md under each folder, along with the requirements.txt and Google Drive links of model checkpoints and processed data.

Contact

For any issues or questions, kindly email us at: Yilun Zhao ([email protected]), or Zhenting Qi ([email protected]).

Citation

@inproceedings{zhao-etal-2023-loft,
    title = "LoFT: Enhancing Faithfulness and Diversity for Table-to-Text Generation via Logic Form Control",
    author = "Zhao, Yilun  and
      Qi, Zhenting  and
      Nan, Linyong  and
      Flores, Lorenzo Jaime and
      Radev, Dragomir",
    booktitle = "Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume",
    month = may,
    year = "2023",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/pdf/2302.02962.pdf"
}

About

Code for EACL 2023 paper "LoFT: Enhancing Faithfulness and Diversity for Table-to-Text Generation via Logic Form Control"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published