Stochastic Wasserstein Gradient Flows using Streaming Data with an Application in Predictive Maintenance
We study estimation problems in safety-critical applications with streaming data. Since estimation problems can be posed as optimization problems in the probability space, we devise a stochastic projected Wasserstein gradient flow that keeps track of the belief of the estimated quantity and can consume samples from online data. We show the convergence properties of our algorithm. Our analysis combines recent advances in the Wasserstein space and its differential structure with more classical stochastic gradient descent. We apply our methodology for predictive maintenance of safety-critical processes: Our approach is shown to lead to superior performance when compared to classical least squares, enabling, among others, improved robustness for decision-making.
Code (0)
등록된 구현이 없습니다.
Tasks
Decision MakingSimilar Papers 제목 키워드 기반
Continuous-time Riemannian SGD and SVRG Flows on Wasserstein Probabilistic Space
Recently, optimization on the Riemannian manifold has provided new insights to the optimization community. In this regard, the manifold taken as the probability measure metric space equipped with the second-order Wassers…
Stochastic OptimizationStreaming Sliced Optimal Transport
Sliced optimal transport (SOT) or sliced Wasserstein (SW) distance is widely recognized for its statistical and computational scalability. In this work, we further enhance the computational scalability by proposing the f…
Change Point DetectionPoint Cloud ClassificationLarge-Scale Wasserstein Gradient Flows
Wasserstein gradient flows provide a powerful means of understanding and solving many diffusion equations. Specifically, Fokker-Planck equations, which model the diffusion of probability measures, can be understood as gr…
A note on convergence of Wasserstein policy optimization
Wasserstein Policy Optimization (WPO) is a recently proposed reinforcement learning algorithm that leverages Wasserstein gradient flows to optimize stochastic policies in continuous action spaces. Despite its empirical s…
Reinforcement LearningA Distributed Algorithm for Measure-valued Optimization with Additive Objective
We propose a distributed nonparametric algorithm for solving measure-valued optimization problems with additive objectives. Such problems arise in several contexts in stochastic learning and control including Langevin sa…