Gaussian process

Related papers: 20

About

A Gaussian process (GP) is a probabilistic machine learning framework that defines a distribution over functions, allowing predictions to be made along with principled uncertainty estimates. Rather than fitting a single deterministic model, a GP treats any finite set of function values as jointly Gaussian-distributed, characterized entirely by a mean function and a covariance (kernel) function that encodes assumptions about smoothness and structure. In robotics and AI, GPs are widely used for learning robot dynamics models, optimizing gaits, planning collision-free trajectories, environmental monitoring with mobile robots, safe reinforcement learning, and shape estimation for grasping. Their ability to quantify prediction uncertainty makes them especially valuable when data is scarce or when safety constraints must be respected — a GP can signal when it is operating outside reliably learned regions. GPs matter because they offer a mathematically rigorous way to learn from limited, noisy data while remaining honest about what is unknown. This uncertainty awareness enables safer exploration, better-informed decision-making, and tighter integration of learned models into control pipelines, making GPs a foundational tool for data-efficient, reliable robot learning.

Top Cited Papers

Learning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models

Seyed Mohammad Khansari-Zadeh, Aude Billard

Citations: 764 • 2011

Gaussian Processes for Data-Efficient Learning in Robotics and Control

Marc Peter Deisenroth, Dieter Fox, Carl Edward Rasmussen

Citations: 656 • 2013

Robot trajectory optimization using approximate inference

Marc Toussaint

Citations: 329 • 2009

Model Learning with Local Gaussian Process Regression

Duy Nguyen-Tuong, Matthias Seeger, Jan Peters

Citations: 322 • 2009

Active global localization for a mobile robot using multiple hypothesis tracking

Patric Jensfelt, Steen Savstrup Kristensen

Citations: 316 • 2001

Most likely heteroscedastic Gaussian process regression

Kristian Kersting, Christian Plagemann, Patrick Pfaff, Wolfram Burgard

Citations: 313 • 2007

Probabilistically safe motion planning to avoid dynamic obstacles with uncertain motion patterns

Georges S. Aoude, Brandon Luders, Joshua Joseph, Nicholas Roy, Jonathan P. How

Citations: 287 • 2013

Reachability-based safe learning with Gaussian processes

Anayo K. Akametalu, Jaime F. Fisac, Jeremy Gillula, Shahab Kaynama, Melanie N. Zeilinger, Claire J. Tomlin

Citations: 262 • 2014

Efficient Informative Sensing using Multiple Robots

Citations: 251 • 2009

Automatic gait optimization with Gaussian process regression

Daniel J. Lizotte, Tao Wang, Michael Bowling, Dale Schuurmans

Citations: 247 • 2007

Manifold Gaussian Processes for regression

Roberto Calandra, Jan Peters, Carl Edward Rasmussen, Marc Peter Deisenroth

Citations: 214 • 2016

Semiparametric Latent Factor Models

Yee Whye Teh, Matthias Seeger, Michael I. Jordan

Citations: 202 • 2005

Gaussian Processes and Reinforcement Learning for Identification and Control of an Autonomous Blimp

Jonathan Ko, Daniel J. Klein, Dieter Fox, Dirk Haehnel

Citations: 199 • 2007

Safe Exploration for Optimization with Gaussian Processes

Yanan Sui, Alkis Gotovos, Joel W. Burdick, Andreas Krause

Citations: 197 • 2015

Learning-based Nonlinear Model Predictive Control to Improve Vision-based Mobile Robot Path Tracking

Chris J. Ostafew, Angela P. Schoellig, Timothy D. Barfoot, Jack Collier

Citations: 195 • 2015

Learning Shared Latent Structure for Image Synthesis and Robotic Imitation

Aaron P. Shon, Keith Grochow, Aaron Hertzmann, Rajesh P. N. Rao

Citations: 183 • 2005

Using model knowledge for learning inverse dynamics

Duy Nguyen-Tuong, Jan Peters

Citations: 176 • 2010

Hybrid Dynamic Moving Obstacle Avoidance Using a Stochastic Reachable Set-Based Potential Field

Nick Malone, Hao-Tien Lewis Chiang, Kendra Lesser, Meeko Oishi, Lydia Tapia

Citations: 169 • 2017

Probabilistic movement modeling for intention inference in human–robot interaction

Zhikun Wang, Katharina Mülling, Marc Peter Deisenroth, Heni Ben Amor, David Vogt, Bernhard Schölkopf, Jan Peters

Citations: 166 • 2013

Robot Programming by Demonstration: A Probabilistic Approach

Sylvain Calinon

Citations: 164 • 2021