Graph Spectral Filtering with Chebyshev Interpolation for Recommendation

Chanwoo Kim, Jinkyu Sung, Yebonn Han, Joonseok Lee

This repository is the official Pytorch implementation for ChebyCF.

Graph convolutional networks have recently gained prominence in collaborative filtering (CF) for recommendations. However, we identify potential bottlenecks in two foundational components. First, the embedding layer leads to a latent space with limited capacity, overlooking locally observed but potentially valuable preference patterns. Also, the widely-used neighborhood aggregation is limited in its ability to leverage diverse preference patterns in a fine-grained manner. Building on spectral graph theory, we reveal that these limitations stem from graph filtering with a cut-off in the frequency spectrum and a restricted linear form. To address these issues, we introduce ChebyCF, a CF framework based on graph spectral filtering. Instead of a learned embedding, it takes a user's raw interaction history to utilize the full spectrum of signals contained in it. Also, it adopts Chebyshev interpolation to effectively approximate a flexible non-linear graph filter, and further enhances it by using an additional ideal pass filter and degree-based normalization. Through extensive experiments, we verify that ChebyCF overcomes the aforementioned bottlenecks and achieves state-of-the-art performance across multiple benchmarks and reasonably fast inference.

Setup

conda create -n chebycf python=3.10
conda activate chebycf
conda install pytorch==1.12.1 cudatoolkit=11.3 -c pytorch
conda install numpy==1.22.4 scipy==1.9.3 pyyaml tqdm

Run

# Execution with command-line arguments
python main.py --model chebycf --dataset last-fm --K 8 --phi 15 --eta 256 --alpha 0.1 --beta 0.2

# Execution with a configuration file
python main.py --config chebycf_last-fm

Performance

Citation

@inproceedings{chanwoo2025chebycf,
  title={Graph Spectral Filtering with Chebyshev Interpolation for Recommendation},
  author={Kim, Chanwoo and Sung, Jinkyu and Han, Yebonn and Lee, Joonseok},
  booktitle={Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval},
  year={2025}
}