Multi-Query Self-Attention with Transition-Aware Embedding Distillation (MQSA-TED)

This is our Tensorflow implementation for the paper:

Tianyu Zhu, Yansong Shi, Yuan Zhang, Yihong Wu, Fengran Mo, and Jian-Yun Nie. "Collaboration and Transition: Distilling Item Transitions into Multi-Query Self-Attention for Sequential Recommendation." WSDM 2024.

Introduction

MQSA-TED is a framework that captures collaborative and transitional signals for sequential recommendation.

Citation

Environment Requirement

The code has been tested running under Python 3.8. The required packages are as follows:

• tensorflow == 2.8.0+
• numpy == 1.23.0+
• scipy == 1.8.0+
• pandas == 1.5.0+

Example to Run the Codes

python main.py --dataset beauty --lr 1e-3 --l2_reg 1e-4 --max_len 50 --dropout_rate 0.5 --L 3 --alpha 0.5 --lambda_kd 0.1 --tau 0.1