MUFFIN: Mixture of User-Adaptive Frequency Filtering for Sequential Recommendation

2025-08-19

AI Summary

Existing frequency-domain sequential recommendation models suffer from two key limitations: incomplete spectral coverage and the absence of personalized filtering. To address these issues, we propose UserAdapt, a novel framework featuring a global–local dual-path frequency-domain modeling architecture integrated with user-adaptive, learnable filters. The global path ensures comprehensive spectral representation across the full frequency range, while the local path selectively enhances behaviorally salient frequency bands. Crucially, filter parameters are dynamically generated from user embeddings, enabling fine-grained, user-specific frequency response control. Extensive experiments on five benchmark datasets demonstrate that UserAdapt consistently outperforms state-of-the-art frequency-domain methods, achieving significant improvements in recommendation accuracy. These results validate its superior capability to capture complex, heterogeneous user behavioral patterns and its robustness across diverse scenarios.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Data Mining & Knowledge Management: Recommender Systems
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Sequential recommendation (SR) aims to predict users' subsequent interactions by modeling their sequential behaviors. Recent studies have explored frequency domain analysis, which effectively models periodic patterns in user sequences. However, existing frequency-domain SR models still face two major drawbacks: (i) limited frequency band coverage, often missing critical behavioral patterns in a specific frequency range, and (ii) lack of personalized frequency filtering, as they apply an identical filter for all users regardless of their distinct frequency characteristics. To address these challenges, we propose a novel frequency-domain model, Mixture of User-adaptive Frequency FIlteriNg (MUFFIN), operating through two complementary modules. (i) The global filtering module (GFM) handles the entire frequency spectrum to capture comprehensive behavioral patterns. (ii) The local filtering module (LFM) selectively emphasizes important frequency bands without excluding information from other ranges. (iii) In both modules, the user-adaptive filter (UAF) is adopted to generate user-specific frequency filters tailored to individual unique characteristics. Finally, by aggregating both modules, MUFFIN captures diverse user behavioral patterns across the full frequency spectrum. Extensive experiments show that MUFFIN consistently outperforms state-of-the-art frequency-domain SR models over five benchmark datasets. The source code is available at https://github.com/ilwoong100/MUFFIN.

Problem

• Addresses limited frequency band coverage in sequential recommendation
• Solves lack of personalized frequency filtering for different users
• Captures comprehensive behavioral patterns across full frequency spectrum

Innovation

• Global filtering module captures entire frequency spectrum
• Local filtering module emphasizes important frequency bands
• User-adaptive filter generates personalized frequency filters