Spectral Stochastic Gaussian Processes for Higher Dimensional Data
Author(s)
Spillman, Matthew David
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
Gaussian Processes are a powerful machine learning algorithm with strong theoretical guarantees. However, full GP inference is computationally infeasible on large or high-dimensional datasets, which has motivated the development of many approximate GP algorithms. We introduce a new approximate Gaussian Process regression algorithm which is able to outperform prior methods on certain high-dimensional and large datasets.
Sponsor
Date
Extent
Resource Type
Text
Resource Subtype
Undergraduate Research Option Thesis