Safe Policy Search with Gaussian Process Models
This work addresses safe policy search for robotics or control systems, but it appears incremental as it builds on the established PILCO framework.
The authors tackled the problem of safely optimizing policy parameters for task execution by training a Gaussian process model based on PILCO to capture system dynamics, enabling closed-form gradient computation and constraint violation probability estimation, with the result that only policies deemed safe are implemented to minimize failure risk.
We propose a method to optimise the parameters of a policy which will be used to safely perform a given task in a data-efficient manner. We train a Gaussian process model to capture the system dynamics, based on the PILCO framework. Our model has useful analytic properties, which allow closed form computation of error gradients and estimating the probability of violating given state space constraints. During training, as well as operation, only policies that are deemed safe are implemented on the real system, minimising the risk of failure.