PETER (PErsonalized Transformer for Explainable Recommendation)

Paper

• Lei Li, Yongfeng Zhang, Li Chen. Personalized Transformer for Explainable Recommendation. ACL'21.

A large pretrained GPT-2 version is available at PEPLER!

A small ecosystem for Recommender Systems-based Natural Language Generation is available at NLG4RS!

Datasets to download

• TripAdvisor Hong Kong
• Amazon Movies & TV
• Yelp 2019

For those who are interested in how to obtain (feature, opinion, template, sentiment) quadruples, please refer to Sentires-Guide.

Usage

Below are examples of how to run PETER (with and without the key feature).

python -u main.py \
--data_path ../TripAdvisor/reviews.pickle \
--index_dir ../TripAdvisor/1/ \
--cuda \
--checkpoint ./tripadvisorf/ \
--peter_mask \
--use_feature >> tripadvisorf.log

python -u main.py \
--data_path ../TripAdvisor/reviews.pickle \
--index_dir ../TripAdvisor/1/ \
--cuda \
--checkpoint ./tripadvisor/ \
--peter_mask >> tripadvisor.log

Code dependencies

• Python 3.6
• PyTorch 1.6

Code references

• Word Language Model
• Sequence-to-Sequence Modeling with nn.Transformer and TorchText
• NLP From Scratch: Translation with a Sequence to Sequence Network and Attention
• Deploying a Seq2Seq Model with TorchScript

Citation

@inproceedings{ACL21-PETER,
	title={Personalized Transformer for Explainable Recommendation},
	author={Li, Lei and Zhang, Yongfeng and Chen, Li},
	booktitle={ACL},
	year={2021}
}