Why not Collaborative Filtering in Dual View? Bridging Sparse and Dense Models

2026-01-14

AI Summary

This work addresses the challenge of modeling user–item interactions in data-sparse scenarios, particularly for cold-item recommendation, where dense embedding models are fundamentally limited by an upper bound on signal-to-noise ratio (SNR). To overcome this limitation, the authors propose the SaD framework, which, for the first time, reveals the complementary nature of sparse and dense models through the lens of SNR theory. SaD introduces a lightweight bidirectional alignment mechanism that synergistically combines the structural reliability of sparse interactions with the semantic expressiveness of dense embeddings, yielding a unified dual-view collaborative filtering model. The framework is plug-and-play and achieves state-of-the-art performance across multiple real-world benchmarks, notably topping the BarsMatch leaderboard, demonstrating that even simple dense models can attain superior results when integrated within the proposed dual-view architecture.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Diffusion Models for Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Collaborative Filtering (CF) remains the cornerstone of modern recommender systems, with dense embedding--based methods dominating current practice. However, these approaches suffer from a critical limitation: our theoretical analysis reveals a fundamental signal-to-noise ratio (SNR) ceiling when modeling unpopular items, where parameter-based dense models experience diminishing SNR under severe data sparsity. To overcome this bottleneck, we propose SaD (Sparse and Dense), a unified framework that integrates the semantic expressiveness of dense embeddings with the structural reliability of sparse interaction patterns. We theoretically show that aligning these dual views yields a strictly superior global SNR. Concretely, SaD introduces a lightweight bidirectional alignment mechanism: the dense view enriches the sparse view by injecting semantic correlations, while the sparse view regularizes the dense model through explicit structural signals. Extensive experiments demonstrate that, under this dual-view alignment, even a simple matrix factorization--style dense model can achieve state-of-the-art performance. Moreover, SaD is plug-and-play and can be seamlessly applied to a wide range of existing recommender models, highlighting the enduring power of collaborative filtering when leveraged from dual perspectives. Further evaluations on real-world benchmarks show that SaD consistently outperforms strong baselines, ranking first on the BarsMatch leaderboard. The code is publicly available at https://github.com/harris26-G/SaD.

Problem

• Collaborative Filtering
• Data Sparsity
• Signal-to-Noise Ratio
• Sparse Interaction
• Dense Embeddings

Innovation

• Collaborative Filtering
• Sparse and Dense Models
• Signal-to-Noise Ratio (SNR)
• Dual-View Alignment
• Recommendation Systems