DeepAI AI Chat
Log In Sign Up

AReS and MaRS - Adversarial and MMD-Minimizing Regression for SDEs

by   Gabriele Abbati, et al.
University of Oxford
Max Planck Society
ETH Zurich

Stochastic differential equations are an important modeling class in many disciplines. Consequently, there exist many methods relying on various discretization and numerical integration schemes. In this paper, we propose a novel, probabilistic model for estimating the drift and diffusion given noisy observations of the underlying stochastic system. Using state-of-the-art adversarial and moment matching inference techniques, we circumvent the use of the discretization schemes as seen in classical approaches. This yields significant improvements in parameter estimation accuracy and robustness given random initial guesses. On four commonly used benchmark systems, we demonstrate the performance of our algorithms compared to state-of-the-art solutions based on extended Kalman filtering and Gaussian processes.


page 1

page 2

page 3

page 4


Lawson schemes for highly oscillatory stochastic differential equations and conservation of invariants

In this paper, we consider a class of stochastic midpoint and trapezoida...

Approximate Bayes learning of stochastic differential equations

We introduce a nonparametric approach for estimating drift and diffusion...

Fenrir: Physics-Enhanced Regression for Initial Value Problems

We show how probabilistic numerics can be used to convert an initial val...

Learning Stochastic Differential Equations With Gaussian Processes Without Gradient Matching

We introduce a novel paradigm for learning non-parametric drift and diff...

Parameter Estimation in Nonlinear Multivariate Stochastic Differential Equations Based on Splitting Schemes

Surprisingly, general estimators for nonlinear continuous time models ba...

Deterministic Inference of Neural Stochastic Differential Equations

Model noise is known to have detrimental effects on neural networks, suc...