Koopman-Equivariant Gaussian Processes
Credible forecasting and representation learning of dynamical systems are of ever-increasing importance for reliable decision-making. To that end, we propose a family of Gaussian processes (GP) for dynamical systems with linear time-invariant responses, which are nonlinear only in initial conditions. This linearity allows us to tractably quantify forecasting and representational uncertainty, simultaneously alleviating the challenge of computing the distribution of trajectories from a GP-based dynamical system and enabling a new probabilistic treatment of learning Koopman operator representations. Using a trajectory-based equivariance -- which we refer to as \textit{Koopman equivariance} -- we obtain a GP model with enhanced generalization capabilities. To allow for large-scale regression, we equip our framework with variational inference based on suitable inducing points. Experiments demonstrate on-par and often better forecasting performance compared to kernel-based methods for learning dynamical systems.
Code (0)
등록된 구현이 없습니다.
Tasks
Decision MakingGaussian ProcessesRepresentation LearningVariational InferenceMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Equivariant Learning of Stochastic Fields: Gaussian Processes and Steerable Conditional Neural Processes
Motivated by objects such as electric fields or fluid streams, we study the problem of learning stochastic fields, i.e. stochastic processes whose samples are fields like those occurring in physics and engineering. Consi…
Gaussian ProcessesTransfer LearningInterrelation of equivariant Gaussian processes and convolutional neural networks
Currently there exists rather promising new trend in machine leaning (ML) based on the relationship between neural networks (NN) and Gaussian processes (GP), including many related subtopics, e.g., signal propagation in …
Gaussian ProcessesKEEC: Koopman Embedded Equivariant Control
An efficient way to control systems with unknown nonlinear dynamics is to find an appropriate embedding or representation for simplified approximation (e.g. linearization), which facilitates system identification and con…
Representation LearningSKooP: Symmetric Koopman Predictions for Faster and More Generalizable Legged Robot Locomotion with Reinforcement Learning
Reinforcement learning (RL) algorithms classically suffer from poor sample efficiency. In robotics, a recent line of work has emerged addressing this problem by encoding physics priors in the learning process. However, m…
Reinforcement LearningInverted Gaussian Process Optimization for Nonparametric Koopman Operator Discovery
The Koopman Operator Theory opens the door for application of rich linear systems theory for computationally efficient modeling and optimal control of nonlinear systems by providing a globally linear representation for c…
GPROperator learningUncertainty Quantification