Beyond Intuition, a Framework for Applying GPs to Real-World Data
Abstract
Gaussian Processes (GPs) offer an attractive method for regression over small, structured and correlated datasets. However, their deployment is hindered by computational costs and limited guidelines on how to apply GPs beyond simple low-dimensional datasets. We propose a framework to identify the suitability of GPs to a given problem and how to set up a robust and well-specified GP model. The guidelines formalise the decisions of experienced GP practitioners, with an emphasis on kernel design and options for computational scalability. The framework is then applied to a case study of glacier elevation change yielding more accurate results at test time.
Cite
@article{arxiv.2307.03093,
title = {Beyond Intuition, a Framework for Applying GPs to Real-World Data},
author = {Kenza Tazi and Jihao Andreas Lin and Ross Viljoen and Alex Gardner and ST John and Hong Ge and Richard E. Turner},
journal= {arXiv preprint arXiv:2307.03093},
year = {2023}
}
Comments
Accepted at the ICML Workshop on Structured Probabilistic Inference and Generative Modelling (2023)