TransIKG (TASLP 2022 Regular Paper)

The code for Exploiting Pairwise Mutual Information for Knowledge-Grounded Dialogue.

Reference

If you use any source code included in this repo in your work, please cite the following paper.

@article{zhang2022exploiting,
  title={Exploiting Pairwise Mutual Information for Knowledge-Grounded Dialogue},
  author={Zhang, Bo and Wang, Jian and Lin, Hongfei and Ma, Hui and Xu, Bo},
  journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing},
  volume={30},
  pages={2231--2240},
  year={2022},
  doi={10.1109/TASLP.2022.3161151}
}

Requirements

• Python 3.8
• Pytorch 1.8
• ParlAI 1.2
• CUDA 11
• Tesla V100 32G

Datasets

We use Wizard of Wikipedia and Holl-E datasets. Note that we used modified verion of Holl-E relased by Kim et al (But they don't release the validation set). Both datasets have already been processed into our defined format, which could be directly download by our code.

Running Codes

Move the folder named parlai_internal to ParlAI.

Wizard of Wikipedia

python parlai_internal/agents/trans_ikg/train_model_wow.py
or
python parlai_internal/agents/trans_ikg_bert/train_model_wow.py

Holl-E

python parlai_internal/agents/trans_ikg/train_model_holle.py
or
python parlai_internal/agents/trans_ikg_bert/train_model_holle.py