Towards Safe Policy Improvement for Non-Stationary MDPs

24 Dec 2020 • blog

How can we ensure in a non-stationary MDP that proposed policy updates provide improvement over an existing policy, with high-confidence? More …

Optimizing for the Future in Non-Stationary MDPs

23 Dec 2020 • blog

How do we create model-free algorithms that can search for a good policy in a non-stationary MDP? More …