Safe reinforcement learning with provable convergence guarantees

EDIC candidacy exam
Exam president: Prof. Patrick Thiran
Thesis advisor: Prof. Maryam Kamgarpour
Co-examiner: Prof. Nicolas Boumal

Abstract
coming soon

Background papers
1. Polyak, B. T. (1963). Gradient methods for the minimisation of functionals. USSR Computational Mathematics and Mathematical Physics, 3(4), 864-878. https://www.sciencedirect.com/science/article/pii/0041555363903823

2. Usmanova, A. Krause, and M. Kamgarpour. Safe non-smooth black-box optimization with application to policy search. In Learning for Dynamics and Control, pages 980–989. PMLR, 2020. https://proceedings.mlr.press/v120/usmanova20a.html

3. Agarwal, A., Kakade, S. M., Lee, J. D., & Mahajan, G. (2021). On the theory of policy gradient methods: Optimality, approximation, and distribution shift. The Journal of Machine Learning Research, 22(1), 4431-4506. https://proceedings.mlr.press/v125/agarwal20a.html