Hypergraph Diffusion for High-Order Recommender Systems

TL;DR

The paper addresses the limitations of existing GNN-based collaborative filtering methods in handling heterophilic user-item interactions and over-smoothing in deep architectures. It introduces WaveHDNN, a two-channel hypergraph diffusion framework comprising a Heterophily-aware Collaborative Encoder and a Multi-scale Group-wise Structure Encoder, coupled with cross-view contrastive learning to align representations. Empirical results on Amazon-Books, Steam, and Yelp show WaveHDNN achieves superior accuracy by capturing high-order group interactions and localized graph structures, with ablations confirming the importance of each component and hyper-parameter studies guiding architectural choices. This work advances high-order recommender systems by integrating wavelet-based diffusion with heterophily-aware messaging, offering a scalable and robust approach to structure-aware recommendation.

Abstract

Recommender systems rely on Collaborative Filtering (CF) to predict user preferences by leveraging patterns in historical user-item interactions. While traditional CF methods primarily focus on learning compact vector embeddings for users and items, graph neural network (GNN)-based approaches have emerged as a powerful alternative, utilizing the structure of user-item interaction graphs to enhance recommendation accuracy. However, existing GNN-based models, such as LightGCN and UltraGCN, often struggle with two major limitations: an inability to fully account for heterophilic interactions, where users engage with diverse item categories, and the over-smoothing problem in multi-layer GNNs, which hinders their ability to model complex, high-order relationships. To address these gaps, we introduce WaveHDNN, an innovative wavelet-enhanced hypergraph diffusion framework. WaveHDNN integrates a Heterophily-aware Collaborative Encoder, designed to capture user-item interactions across diverse categories, with a Multi-scale Group-wise Structure Encoder, which leverages wavelet transforms to effectively model localized graph structures. Additionally, cross-view contrastive learning is employed to maintain robust and consistent representations. Experiments on benchmark datasets validate the efficacy of WaveHDNN, demonstrating its superior ability to capture both heterophilic and localized structural information, leading to improved recommendation performance.

Figures (3)

• Figure 1: An illustration of WaveHDNN.
• Figure 2: Effects of no. HGCN layers.
• Figure 3: Effects of embedding dimensions.