IDURL

The source code of our SIGIR 2025 paper "Towards Interest Drift-driven User Representation Learning in Sequential Recommendation"

Preparation

We train and evaluate our IDURL using a Tesla V100 GPU with 32 GB memory. <br> Our code requires the following python packages:

• numpy==1.21.6
• scipy==1.7.3
• torch==1.13.1+cu116
• tensorboard==2.11.2

Usage

We provide two datasets, i.e., Beauty ./dataset/Amazon_Beauty and Toys ./dataset/Amazon_Toys_and_Games. <br> Please download the other two datasets from RecSysDatasets or their Google Drive. And put the files in ./dataset/ like the following.

$ tree
.
├── Amazon_Beauty
│   ├── Amazon_Beauty.inter
│   └── Amazon_Beauty.item
├── Amazon_Toys_and_Games
│   ├── Amazon_Toys_and_Games.inter
│   └── Amazon_Toys_and_Games.item
├── Amazon_Sports_and_Outdoors
│   ├── Amazon_Sports_and_Outdoors.inter
│   └── Amazon_Sports_and_Outdoors.item
└── yelp
    ├── README.md
    ├── yelp.inter
    ├── yelp.item
    └── yelp.user

Run the command./run_IDURL.sh. After training and evaluation, check out the results in ./run_results/.

Contact

If you have any questions, please send emails to Xiaolin Lin (linxiaolin2021@email.szu.edu.com).

Credit

This repository is based on RecBole and DIF-SR.

Citation

@inproceedings{10.1145/3726302.3730099,
author = {Lin, Xiaolin and Pan, Weike and Ming, Zhong},
title = {Towards Interest Drift-driven User Representation Learning in Sequential Recommendation},
year = {2025},
url = {https://doi.org/10.1145/3726302.3730099},
doi = {10.1145/3726302.3730099},
booktitle = {Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval},
pages = {1541–1551}
}