Revisiting Active Sets for Gaussian Process Decoders
Abstract
Decoders built on Gaussian processes (GPs) are enticing due to the marginalisation over the non-linear function space. Such models (also known as GP-LVMs) are often expensive and notoriously difficult to train in practice, but can be scaled using variational inference and inducing points. In this paper, we revisit active set approximations. We develop a new stochastic estimate of the log-marginal likelihood based on recently discovered links to cross-validation, and propose a computationally efficient approximation thereof. We demonstrate that the resulting stochastic active sets (SAS) approximation significantly improves the robustness of GP decoder training while reducing computational cost. The SAS-GP obtains more structure in the latent space, scales to many datapoints and learns better representations than variational autoencoders, which is rarely the case for GP decoders.
Cite
@article{arxiv.2209.04636,
title = {Revisiting Active Sets for Gaussian Process Decoders},
author = {Pablo Moreno-Muñoz and Cilie W Feldager and Søren Hauberg},
journal= {arXiv preprint arXiv:2209.04636},
year = {2022}
}
Comments
Accepted at Advances in Neural Information Processing Systems (NeurIPS) 2022