AISTATS 2021poster41 citations

Kernel Interpolation for Scalable Online Gaussian Processes

Samuel Stanton, Wesley Maddox, Ian Delbridge, Andrew Gordon Wilson

Abstract

Gaussian processes (GPs) provide a gold standard for performance in online settings, such as sample-efficient control and black box optimization, where we need to update a posterior distribution as we acquire data in a sequential online setting. However, updating a GP posterior to accommodate even a single new observation after having observed $n$ points incurs at least $\mathcal{O}(n)$ computations in the exact setting. We show how to use structured kernel interpolation to efficiently reuse computations for constant-time $\mathcal{O}(1)$ online updates with respect to the number of points $n$, while retaining exact inference. We demonstrate the promise of our approach in a range of online regression and classification settings, Bayesian optimization, and active sampling to reduce error in malaria incidence forecasting.

BibTeX
@InProceedings{pmlr-v130-stanton21a,
  title = 	 { Kernel Interpolation for Scalable Online Gaussian Processes },
  author =       {Stanton, Samuel and Maddox, Wesley and Delbridge, Ian and Gordon Wilson, Andrew},
  booktitle = 	 {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics},
  pages = 	 {3133--3141},
  year = 	 {2021},
  editor = 	 {Banerjee, Arindam and Fukumizu, Kenji},
  volume = 	 {130},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {13--15 Apr},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v130/stanton21a/stanton21a.pdf},
  url = 	 {https://proceedings.mlr.press/v130/stanton21a.html},
  abstract = 	 { Gaussian processes (GPs) provide a gold standard for performance in online settings, such as sample-efficient control and black box optimization, where we need to update a posterior distribution as we acquire data in a sequential online setting. However, updating a GP posterior to accommodate even a single new observation after having observed $n$ points incurs at least $\mathcal{O}(n)$ computations in the exact setting. We show how to use structured kernel interpolation to efficiently reuse computations for constant-time $\mathcal{O}(1)$ online updates with respect to the number of points $n$, while retaining exact inference. We demonstrate the promise of our approach in a range of online regression and classification settings, Bayesian optimization, and active sampling to reduce error in malaria incidence forecasting. }
}
Kernel Interpolation for Scalable Online Gaussian Processes · AISTATS 2021