TimeMM: Time-as-Operator Spectral Filtering for Dynamic Multimodal Recommendation

2026-04-28

AI Summary

Existing approaches struggle to effectively model the continuous and non-stationary evolution of user interests in multimodal recommendation and lack fine-grained adaptation to the distinct temporal sensitivities of visual and textual modalities across varying time scales. To address these limitations, this work proposes TimeMM, a novel framework that treats time as a spectral-domain operator acting on the user-item interaction graph. TimeMM dynamically adjusts edge weights through time-conditioned spectral filtering without explicit eigendecomposition and integrates adaptive operator mixing, spectrum-aware modality routing, and spectral diversity regularization to jointly capture non-stationary interest dynamics and modality-specific temporal sensitivities. The method achieves significant performance gains over state-of-the-art models on multiple real-world benchmarks while maintaining linear time complexity.

Abstract

Multimodal recommendation improves user modeling by integrating collaborative signals with heterogeneous item content. In real applications, user interests evolve over time and exhibit nonstationary dynamics, where different preference factors change at different rates. This challenge is amplified in multimodal settings because visual and textual cues can dominate decisions under different temporal regimes. Despite strong progress, most multimodal recommenders still rely on static interaction graphs or coarse temporal heuristics, which limits their ability to model continuous preference evolution with fine-grained temporal adaptation. To address these limitations, we propose TimeMM, a time-conditioned spectral filtering framework for dynamic multimodal recommendation. TimeMM instantiates Time-as-Operator by mapping interaction recency to a family of parametric temporal kernels that reweight edges on the user--item graph, producing component-specific representations without explicit eigendecomposition. To capture non-stationary interests, we introduce Adaptive Spectral Filtering that mixes the operator bank according to temporal context, yielding prediction-specific effective spectral responses. To account for modality-specific temporal sensitivity, we further propose Spectral-Aware Modality Routing that calibrates visual and textual contributions conditioned on the same temporal context. Finally, a ranking-space Spectral Diversity Regularization encourages complementary expert behaviors and prevents filter-bank collapse. Extensive experiments on real-world benchmarks demonstrate that TimeMM consistently outperforms state-of-the-art multimodal recommenders while maintaining linear-time scalability.

Problem

• multimodal recommendation
• temporal dynamics
• nonstationary user interests
• time-aware modeling
• spectral filtering

Innovation

• Time-as-Operator
• Adaptive Spectral Filtering
• Spectral-Aware Modality Routing
• Nonstationary Dynamics
• Dynamic Multimodal Recommendation