Variational Gaussian Process

Variational Gaussian Processes (VGPs) are approximate inference methods designed to address the computational challenges of using Gaussian Processes (GPs) for large datasets or high-dimensional problems. Current research focuses on improving the scalability and accuracy of VGPs through techniques like sparse approximations using inducing points, coresets, and novel variational families, as well as optimizing inference algorithms such as stochastic gradient descent and natural gradient descent. These advancements enable the application of VGPs to diverse fields, including Bayesian optimization, spatial prediction, and robot motion planning, where accurate and efficient probabilistic modeling is crucial. The resulting improvements in computational efficiency and predictive accuracy are significantly impacting various scientific and engineering domains.

Papers

June 6, 2024

Approximation-Aware Bayesian Optimization
Natalie Maus, Kyurae Kim, Geoff Pleiss, David Eriksson, John P. Cunningham, Jacob R. Gardner
Bayesian Optimization Variational Gaussian Process Utility Learning

February 23, 2024

Streaming Gaussian Dirichlet Random Fields for Spatial Predictions of High Dimensional Categorical Observations
J. E. San Soucie, H. M. Sosik, Y. Girdhar
Fast Inference SpatioTemporal Data Spatial Prediction Variational Gaussian Process

November 2, 2023

A Coreset-based, Tempered Variational Posterior for Accurate and Scalable Stochastic Gaussian Process Inference
Mert Ketenci, Adler Perotte, Noémie Elhadad, Iñigo Urteaga
Gaussian Process COreset Selection Variational Posterior Marginal Likelihood Coreset Size Variational Family Variational Gaussian Process GP Inference

November 1, 2023

Robust and Conjugate Gaussian Process Regression
Matias Altamirano, François-Xavier Briol, Jeremias Knoblauch
Gaussian Process Robust Version Variational Gaussian Process Implicit Conditioning Method Generalised Bayesian

October 31, 2023

Stochastic Gradient Descent for Gaussian Processes Done Right
Jihao Andreas Lin, Shreyas Padhy, Javier Antorán, Austin Tripp, Alexander Terenin, Csaba Szepesvári, José Miguel Hernández-Lobato, David Janz
Gaussian Process Stochastic Gradient Descent Gaussian Process Regression Bayesian Optimisation Variational Gaussian Process

September 2, 2023

A Unifying Variational Framework for Gaussian Process Motion Planning
Lucas Cosier, Rares Iordan, Sicelukwanda Zwane, Giovanni Franzese, James T. Wilson, Marc Peter Deisenroth, Alexander Terenin, Yasemin Bekiroglu
Variational Method Robot Motion Planning Motion Planning Algorithm Soft Constraint Variational Gaussian Process

June 3, 2023

Variational Gaussian Process Diffusion Processes
Prakhar Verma, Vincent Adam, Arno Solin
Variational Inference Stochastic Differential Equation Variational Gaussian Process Continual Diffusion Linear Diffusion

April 27, 2023

Spherical Inducing Features for Orthogonally-Decoupled Gaussian Processes
Louis C. Tiao, Vincent Dutordoir, Victor Picheny
Neural Network Gaussian Process Feature Wise Spherical Surface Basis Function Variational Gaussian Process

July 16, 2022

Learning inducing points and uncertainty on molecular data by scalable variational Gaussian processes
Mikhail Tsitsvero, Mingoo Jin, Andrey Lyalin
LeArning Abstract High Uncertainty Anticipation Exceptional Point Molecular Dynamic Molecular Data Variational Distribution Variational Gaussian Process De Novo Crystal

March 21, 2022

Fully-probabilistic Terrain Modelling with Stochastic Variational Gaussian Process Maps
Ignacio Torroba, Christopher Illife Sprague, John Folkesson
Gaussian Process Uncertainty Map Variational Gaussian Process Terrain Model P Map

March 14, 2022

Non-Parametric Modeling of Spatio-Temporal Human Activity Based on Mobile Robot Observations
Marvin Stuede, Moritz Schappler
Spatio Temporal Autonomous Robot Predictive Uncertainty Variational Gaussian Process

February 3, 2022

Variational Nearest Neighbor Gaussian Process
Luhuan Wu, Geoff Pleiss, John Cunningham
Gaussian Process Variational Method Variational Approximation Sparse Approximation Variational Gaussian Process

November 5, 2021

Dual Parameterization of Sparse Variational Gaussian Processes
Vincent Adam, Paul E. Chang, Mohammad Emtiyaz Khan, Arno Solin
Sparse Variational Gaussian Process Variational Gaussian Process