Learning and Policy Search in Stochastic Dynamical Systems with Bayesian Neural Networks

by   Stefan Depeweg, et al.
University of Cambridge
Siemens AG
Harvard University

We present an algorithm for model-based reinforcement learning that combines Bayesian neural networks (BNNs) with random roll-outs and stochastic optimization for policy learning. The BNNs are trained by minimizing α-divergences, allowing us to capture complicated statistical patterns in the transition dynamics, e.g. multi-modality and heteroskedasticity, which are usually missed by other common modeling approaches. We illustrate the performance of our method by solving a challenging benchmark where model-based approaches usually fail and by obtaining promising results in a real-world scenario for controlling a gas turbine.


page 6

page 12


Augmented Random Search for Quadcopter Control: An alternative to Reinforcement Learning

Model-based reinforcement learning strategies are believed to exhibit mo...

Bootstrapping the Expressivity with Model-based Planning

We compare the model-free reinforcement learning with the model-based ap...

Gradient-Aware Model-based Policy Search

Traditional model-based reinforcement learning approaches learn a model ...

SpReME: Sparse Regression for Multi-Environment Dynamic Systems

Learning dynamical systems is a promising avenue for scientific discover...

Automatic detection of alarm sounds in a noisy hospital environment using model and non-model based approaches

In the noisy acoustic environment of a Neonatal Intensive Care Unit (NIC...

Random Actions vs Random Policies: Bootstrapping Model-Based Direct Policy Search

This paper studies the impact of the initial data gathering method on th...

Probabilistic Evolution of Stochastic Dynamical Systems: A Meso-scale Perspective

Stochastic dynamical systems arise naturally across nearly all areas of ...

Please sign up or login with your details

Forgot password? Click here to reset