Sparse GP

Sparse Gaussian Processes (GPs) aim to overcome the computational limitations of standard GPs, which scale poorly with large datasets, while preserving their desirable properties like uncertainty quantification. Current research focuses on developing efficient algorithms, such as those employing Kronecker products, binary tree kernels, and variational inference with normalizing flows, to achieve linear or near-linear time complexity. These advancements enable the application of GPs to significantly larger datasets in diverse fields, including fluid mechanics, robotics, and 3D scene reconstruction, improving model accuracy and uncertainty estimation in computationally demanding applications.

Papers

December 13, 2023

A Kronecker product accelerated efficient sparse Gaussian Process (E-SGP) for flow emulation
Yu Duan, Matthew Eaton, Michael Bluck
Sparse Gaussian Process Kronecker Product Surrogate Modelling Sparse GP

November 30, 2023

SparseGS: Real-Time 360{\deg} Sparse View Synthesis using Gaussian Splatting
Haolin Xiong, Sairisheek Muttukuru, Rishi Upadhyay, Pradyumna Chari, Achuta Kadambi
Neural Radiance Field Gaussian Splatting Novel View Synthesis View Synthesis Implicit Scene Sparse GP

January 21, 2023

Towards Flexibility and Interpretability of Gaussian Process State-Space Model
Zhid Lin, Feng Yin, Juan Maroñas
Inherent Interpretability Gaussian Process State Space Model State Space Scalable Variational Sparse GP

October 4, 2022

Log-Linear-Time Gaussian Processes Using Binary Tree Kernels
Michael K. Cohen, Samuel Daulton, Michael A. Osborne
Gaussian Process Gaussian Process Regression Sparse Kernel Sparse GP Subtree Kernel

July 30, 2022

Learning robot inverse dynamics using sparse online Gaussian process with forgetting mechanism
Wei Li, Zhiwen Li, Yiqi Liu, Yongping Pan
Functional Mechanism Sparse Gaussian Process Sparse Approximation Inverse Dynamic Sparse GP

May 18, 2022

Exact Gaussian Processes for Massive Datasets via Non-Stationary Sparsity-Discovering Kernels
Marcus M. Noack, Harinarayan Krishnan, Mark D. Risser, Kristofer G. Reyes
Gaussian Process Non Stationary Stochastic Approximation Small Kernel Sparse Structure Stationary Kernel Sparse GP

December 17, 2021

Correlated Product of Experts for Sparse Gaussian Process Regression
Manuel Schürch, Dario Azzimonti, Alessio Benavoli, Marco Zaffalon
Gaussian Process Expert Knowledge Regression Task Correlated Source Sparse GP Sparse Gaussian Process Regression GP Inference