ContextualSP/logic_gan

LogiGAN

This repository serves primarily as codebase and data, model for training, evaluation and inference of the logical pre-training method LogiGAN. LogiGAN (NeurIPS 2022) is the adversarial logical pre-training method with Transformer-based encoder-decoder backbone.

Preprocessing

Citation

If you find this resource useful, please cite the paper introducing CARP:

@article{pi2022logigan,
  title={LogiGAN: Learning Logical Reasoning via Adversarial Pre-training},
  author={Pi, Xinyu and Zhong, Wanjun and Gao, Yan and Duan, Nan and Lou, Jian-Guang},
  journal={arXiv preprint arXiv:2205.08794},
  year={2022}
}