When Bayesian Tensor Completion Meets Multioutput Gaussian Processes: Functional Universality and Rank Learning
- 2 December 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 73 (1053587X) , 5319-5335
Abstract
Functional tensor decomposition can analyze multidimensional data with real-valued indices, paving the path for applications in machine learning and signal processing. A limitation of existing approaches is the assumption that the tensor rank—a critical parameter governing model complexity—is known. However, determining the optimal rank is a nondeterministic polynomial-time hard (NP-hard) task and there is a limited understanding regarding the expressive power of functional low-rank tensor models for continuous signals.We propose a rank-revealing functional Bayesian tensor completion (RRFBTC) method. Modeling the latent functions through carefully designed multioutput Gaussian processes, RR-FBTC handles tensors with real-valued indices while enabling automatic tensor rank determination during the inference process. We establish the universal approximation property of the model for continuous multi-dimensional signals, demonstrating its expressive power in a concise format. To learn this model, we employ the variational inference framework and derive an efficient algorithm with closed-form updates. Experiments on both synthetic and realworld datasets demonstrate the effectiveness and superiority of the RR-FBTC over state-of-the-art approaches.Keywords
Funding Information
- National Natural Science Foundation of China (62371418)
- Fundamental Research Funds for the Central Universities (226-2025-00168)
- Shenzhen Science and Technology Program (JCYJ20220530143806016)
- European Commission (101136568 - HERON)
This publication has 48 references indexed in Scilit:
- Tensor train factorization under noisy and incomplete data with automatic rank estimationPattern Recognition, 2023
- Multivariate Gaussian and Student-t process regression for multi-output predictionNeural Computing and Applications, 2019
- Shrinkage priors for Bayesian penalized regressionJournal of Mathematical Psychology, 2019
- Support Vector Machines2008
- A new efficient method for determining the number of components in PARAFAC modelsJournal of Chemometrics, 2003
- On Uniqueness in Candecomp/ParafacPsychometrika, 2002
- Priors for Infinite Networks1996
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989
- Some Mathematical Notes on Three-Mode Factor AnalysisPsychometrika, 1966
- The Expression of a Tensor or a Polyadic as a Sum of ProductsJournal of Mathematics and Physics, 1927