Gaussian Processes with Bayesian Inference of Covariate Couplings
Abstract
Gaussian processes are powerful probabilistic models that are often coupled with ARD capable of uncovering the importance of individual covariates. We develop covariances characterized by affine transformations of the inputs, formalized via a precision matrix between covariates, which can uncover covariate couplings for enhanced interpretability. We study a range of couplings priors from Wishart to Horseshoe and present fully Bayesian inference of such precision matrices within sparse Gaussian processes. We empirically demonstrate the efficacy and interpretability of this approach.
Cite
Text
Rosso et al. "Gaussian Processes with Bayesian Inference of Covariate Couplings." Transactions on Machine Learning Research, 2025.Markdown
[Rosso et al. "Gaussian Processes with Bayesian Inference of Covariate Couplings." Transactions on Machine Learning Research, 2025.](https://mlanthology.org/tmlr/2025/rosso2025tmlr-gaussian/)BibTeX
@article{rosso2025tmlr-gaussian,
title = {{Gaussian Processes with Bayesian Inference of Covariate Couplings}},
author = {Rosso, Mattia and Ylä-Jääski, Juho and Shen, Zheyang and Heinonen, Markus and Filippone, Maurizio},
journal = {Transactions on Machine Learning Research},
year = {2025},
url = {https://mlanthology.org/tmlr/2025/rosso2025tmlr-gaussian/}
}