English

Variational sparse inverse Cholesky approximation for latent Gaussian processes via double Kullback-Leibler minimization

Machine Learning 2023-05-30 v2 Machine Learning Computation

Abstract

To achieve scalable and accurate inference for latent Gaussian processes, we propose a variational approximation based on a family of Gaussian distributions whose covariance matrices have sparse inverse Cholesky (SIC) factors. We combine this variational approximation of the posterior with a similar and efficient SIC-restricted Kullback-Leibler-optimal approximation of the prior. We then focus on a particular SIC ordering and nearest-neighbor-based sparsity pattern resulting in highly accurate prior and posterior approximations. For this setting, our variational approximation can be computed via stochastic gradient descent in polylogarithmic time per iteration. We provide numerical comparisons showing that the proposed double-Kullback-Leibler-optimal Gaussian-process approximation (DKLGP) can sometimes be vastly more accurate for stationary kernels than alternative approaches such as inducing-point and mean-field approximations at similar computational complexity.

Keywords

Cite

@article{arxiv.2301.13303,
  title  = {Variational sparse inverse Cholesky approximation for latent Gaussian processes via double Kullback-Leibler minimization},
  author = {Jian Cao and Myeongjong Kang and Felix Jimenez and Huiyan Sang and Florian Schafer and Matthias Katzfuss},
  journal= {arXiv preprint arXiv:2301.13303},
  year   = {2023}
}

Comments

Accepted at the 2023 International Conference on Machine Learning (ICML). 18 pages with references and appendices, 14 figures

R2 v1 2026-06-28T08:27:29.329Z