RecSys'20 TAFA: Two-headed Attention Fused Autoencoder for Context-Aware Recommendations

Authors: Jinpeng Zhou*, Zhaoyue Cheng*, Felipe Perez, Maksims Volkovs
[paper]

Environment:

The code was developed and tested on the following python environment:

python 3.7.6
pytorch 1.4.0
pandas 1.1.0
tqdm
allennlp
gensim
cupy
nltk

Instructions:

To train and eveluate TAFA on the Amazon datasets (digital music, grocery and gourmet food, video games, cds and vinyl), run this command:

sh run_amazon.sh [digital_music|grocery_and_gourmet_food|video_games|cds_and_vinyl]

Citation

If you find this code useful in your research, please cite the following paper:

@inproceedings{zhou2020tafa,
  title={TAFA: {Two-headed} Attention Fused Autoencoder for Context-Aware Recommendations},
  author={Jinpeng Zhou, Zhaoyue Cheng, Felipe Perez, Maksims Volkovs},
  booktitle={RecSys},
  year={2020}
}