T3SRS: Tensor Train Transformer for compressing sequential recommender systems

Hao Li, Jianli Zhao, Huan Huo, Sheng Fang, Jianjian Chen, Lutong Yao, Yiran Hua

Abstract

Highlights•A novel model as a remedy for parameter redundancy in SRSs is proposed.•A tensor train layer has lower space complexity than the original linear layer.•The proposed model compresses SRSs parameters effectively.•Tensor Train Transformer model is broadly applicable.
BibTeX
@article{li2024t,
  author = {Hao Li and Jianli Zhao and Huan Huo and Sheng Fang and Jianjian Chen and Lutong Yao and Yiran Hua},
  title = {{T3SRS: Tensor Train Transformer for compressing sequential recommender systems}},
  journal = {Expert Systems with Applications},
  year = {2024},
  url = {https://openreview.net/forum?id=32F2YbPtW4},
}

Loading related content...

Figure