Probabilistic Residual User Clustering

TMLR Paper5778 Authors

Abstract: Modern recommender systems are typically based on deep learning (DL) models, where a dense encoder learns representations of users and items. As a result, these systems often suffer from the black-box nature and computational complexity of the underlying models, making it difficult to systematically interpret their outputs and enhance their recommendation capabilities. To address this problem, we propose Probabilistic Residual User Clustering (PRUC), a causal Bayesian recommendation model based on user clustering. Specifically, we address this problem by (1) dividing users into clusters in an unsupervised manner and identifying causal confounders that influence latent variables, (2) developing sub-models for each confounder given the observable variables, and (3) generating recommendations by aggregating the rating residuals under each confounder using do-calculus. Experiments demonstrate that our plug-and-play PRUC is compatible with various base DL recommender systems, significantly improving their performance while automatically discovering meaningful user clusters.

Submission Length: Regular submission (no more than 12 pages of main content)

Previous TMLR Submission Url: https://openreview.net/forum?id=5KeMRVIZpH

Changes Since Last Submission: We have restored the font.

Assigned Action Editor: ~Seungjin_Choi1

Submission Number: 5778