Propagating Similarity, Mitigating Uncertainty: Similarity Propagation-enhanced Uncertainty for Multimodal Recommendation

2026-01-27

AI Summary

This work addresses the limitations of existing multimodal recommendation systems, which are often hindered by modality-specific noise and uncertainty that are typically overlooked, leading to suboptimal fusion performance. To tackle this issue, the authors propose an uncertainty-aware multimodal recommendation framework that constructs both modality similarity graphs and collaborative similarity graphs to refine user and item representations from dual perspectives—content and behavior. A novel adaptive preference aggregation module is introduced to dynamically integrate more reliable modality features based on their estimated uncertainties. Notably, this is the first approach to jointly model similarity propagation and modality-specific uncertainty, substantially enhancing representation quality and fusion efficiency. Extensive experiments on three benchmark datasets demonstrate that the proposed method significantly outperforms state-of-the-art models, validating its effectiveness and innovation.

Technology Category

• Machine Learning: Multimodal Learning
• Reasoning under Uncertainty: Uncertainty Representations
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Multimodal Recommendation (MMR) systems are crucial for modern platforms but are often hampered by inherent noise and uncertainty in modal features, such as blurry images, diverse visual appearances, or ambiguous text. Existing methods often overlook this modality-specific uncertainty, leading to ineffective feature fusion. Furthermore, they fail to leverage rich similarity patterns among users and items to refine representations and their corresponding uncertainty estimates. To address these challenges, we propose a novel framework, Similarity Propagation-enhanced Uncertainty for Multimodal Recommendation (SPUMR). SPUMR explicitly models and mitigates uncertainty by first constructing the Modality Similarity Graph and the Collaborative Similarity Graph to refine representations from both content and behavioral perspectives. The Uncertainty-aware Preference Aggregation module then adaptively fuses the refined multimodal features, assigning greater weight to more reliable modalities. Extensive experiments on three benchmark datasets demonstrate that SPUMR achieves significant improvements over existing leading methods.

Problem

• Multimodal Recommendation
• Uncertainty
• Similarity Propagation
• Feature Fusion
• Modality Noise

Innovation

• Similarity Propagation
• Uncertainty Modeling
• Multimodal Recommendation
• Graph-based Representation
• Adaptive Fusion