IGDMRec: Behavior Conditioned Item Graph Diffusion for Multimodal Recommendation

2025-12-22

AI Summary

In multimodal recommendation, semantic item graphs are vulnerable to modality noise and behavior–semantics misalignment, leading to spurious edges and biased recommendations. To address this, we propose Behavior-Guided Diffusion Denoising (BGD), a novel framework that introduces a behavior-conditioned graph diffusion mechanism and classifier-free guidance for diffusion modeling, explicitly injecting user interaction signals into graph structure learning. BGD further integrates contrastive representation enhancement to jointly optimize semantic and behavioral graphs. The method employs a lightweight Conditional Denoising Network (CD-Net), balancing computational efficiency with expressive power. Extensive experiments on four real-world datasets demonstrate significant improvements over state-of-the-art methods. Ablation studies confirm the necessity of each component, while robustness analysis shows strong resilience against diverse noise types.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Computer Vision: Diffusion Models for Vision
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Multimodal recommender systems (MRSs) are critical for various online platforms, offering users more accurate personalized recommendations by incorporating multimodal information of items. Structure-based MRSs have achieved state-of-the-art performance by constructing semantic item graphs, which explicitly model relationships between items based on modality feature similarity. However, such semantic item graphs are often noisy due to 1) inherent noise in multimodal information and 2) misalignment between item semantics and user-item co-occurrence relationships, which introduces false links and leads to suboptimal recommendations. To address this challenge, we propose Item Graph Diffusion for Multimodal Recommendation (IGDMRec), a novel method that leverages a diffusion model with classifier-free guidance to denoise the semantic item graph by integrating user behavioral information. Specifically, IGDMRec introduces a Behavior-conditioned Graph Diffusion (BGD) module, incorporating interaction data as conditioning information to guide the denoising of the semantic item graph. Additionally, a Conditional Denoising Network (CD-Net) is designed to implement the denoising process with manageable complexity. Finally, we propose a contrastive representation augmentation scheme that leverages both the denoised item graph and the original item graph to enhance item representations. LL{Extensive experiments on four real-world datasets demonstrate the superiority of IGDMRec over competitive baselines, with robustness analysis validating its denoising capability and ablation studies verifying the effectiveness of its key components.

Problem

• Denoise semantic item graphs in multimodal recommendation
• Align item semantics with user behavior patterns
• Enhance item representations using denoised graphs

Innovation

• Leverages diffusion model with classifier-free guidance for denoising
• Incorporates user behavior to condition graph diffusion process
• Uses contrastive augmentation with denoised and original graphs