Collaborative filtering based on nonnegative/binary matrix factorization

2024-10-14

arXiv.org

AI Summary

To address low prediction accuracy in collaborative filtering caused by highly sparse user rating data, this paper proposes Masked Non-negative/Binary Matrix Factorization (NBMF). It is the first work to adapt NBMF to sparse recommendation settings by explicitly masking unobserved entries to formulate a more robust objective function and leveraging low-latency Ising hardware for accelerated optimization. Key contributions include: (1) a sparsity-aware masking mechanism that eliminates interference from unobserved interactions during factorization; (2) joint integration of non-negativity and binary constraints to enhance model interpretability and generalization; and (3) end-to-end hardware–algorithm co-optimization. Experiments on multiple sparse benchmark datasets demonstrate that the proposed method significantly outperforms conventional NMF and state-of-the-art collaborative filtering approaches, achieving an average 8.3% improvement in prediction accuracy and a 42% reduction in inference latency.

Technology Category

Machine Learning: Hardware-aware MLData Mining & Knowledge Management: Recommender SystemsCognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendationGraph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphsSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Collaborative filtering generates recommendations based on user-item similarities through rating data, which may involve numerous unrated items. To predict scores for unrated items, matrix factorization techniques, such as nonnegative matrix factorization (NMF), are often employed to predict scores for unrated items. Nonnegative/binary matrix factorization (NBMF), which is an extension of NMF, approximates a nonnegative matrix as the product of nonnegative and binary matrices. Previous studies have employed NBMF for image analysis where the data were dense. In this paper, we propose a modified NBMF algorithm that can be applied to collaborative filtering where data are sparse. In the modified method, unrated elements in a rating matrix are masked, which improves the collaborative filtering performance. Utilizing a low-latency Ising machine in NBMF is advantageous in terms of the computation time, making the proposed method beneficial.

Problem

• Sparse Data
• Collaborative Filtering
• Recommendation Systems

Innovation

• Non-negative Binary Matrix Factorization
• Sparse Data Optimization
• Low-latency Ising Machine