DeepAI AI Chat
Log In Sign Up

Non-Factorised Variational Inference in Dynamical Systems

by   Alessandro Davide Ialongo, et al.
University of Cambridge

We focus on variational inference in dynamical systems where the discrete time transition function (or evolution rule) is modelled by a Gaussian process. The dominant approach so far has been to use a factorised posterior distribution, decoupling the transition function from the system states. This is not exact in general and can lead to an overconfident posterior over the transition function as well as an overestimation of the intrinsic stochasticity of the system (process noise). We propose a new method that addresses these issues and incurs no additional computational costs.


page 1

page 2

page 3

page 4


Overcoming Mean-Field Approximations in Recurrent Gaussian Process Models

We identify a new variational inference scheme for dynamical systems who...

Online Gaussian Process State-Space Models: Learning and Planning for Partially Observable Dynamical Systems

Gaussian process state-space model (GPSSM) is a probabilistic dynamical ...

Variational Inference for Continuous-Time Switching Dynamical Systems

Switching dynamical systems provide a powerful, interpretable modeling f...

A Large Deviation Approach to Posterior Consistency in Dynamical Systems

In this paper, we provide asymptotic results concerning (generalized) Ba...

The Most Likely Transition Path for a Class of Distribution-Dependent Stochastic Systems

Distribution-dependent stochastic dynamical systems arise widely in engi...

An Optimal Control Method to Compute the Most Likely Transition Path for Stochastic Dynamical Systems with Jumps

Many complex real world phenomena exhibit abrupt, intermittent or jumpin...

Profiles of dynamical systems and their algebra

The commutative semiring 𝐃 of finite, discrete-time dynamical systems wa...