If you are encountering a collapse of the policy, one of the general approaches I have found helpful is to increase the number of experiences in the training batch. I am currently using ~100 for my experiments. In my experience this provides a more robust value estimate, and fewer policy collapses.

PhD. Interests include Deep (Reinforcement) Learning, Computational Neuroscience, and Phenomenology.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store