Bayesian high-order tensor factorization for learning the hidden low-rank structure

Jianan Liu

doi:10.1016/j.patcog.2025.112608

Bayesian high-order tensor factorization for learning the hidden low-rank structure

2025-10-15

0

PRE

AI

DOI：10.1016/j.patcog.2025.112608

Original

Original request for help

Save

Abstract

En 中文

• We propose a Bayesian High-order Tensor Factorization (BHTF) framework within which we introduce proper priors and explicitly model the uncertainty to accurately learn the hidden low-rank structure from a high-order tensor observation. • The proposed BHTF can automatically determine the rank of the low-rank structure. • We specify the BHTF in both the original and transform domain, and establish a connection between the two domains. • We propose a variational inference algorithm that crosses the original domain and the transform domain for efficient parameters estimation.

Bayesian high-order tensor factorization for learning the hidden low-rank structure

Abstract

Journal

Organization