Rahul Savani

The Complexity of All-switches Strategy Improvement

2015

Strategy improvement is a widely-used and well-studied class of algorithms for solving graph-based infinite games. These algorithms are parametrized by a switching rule, and one of the most natural rules is “all switches” which switches …

A Natural Actor-Critic Algorithm with Downside Risk Constraints

2020 · arXiv (Cornell University)

Existing work on risk-sensitive reinforcement learning - both for symmetric and downside risk measures - has typically used direct Monte-Carlo estimation of policy gradients. While this approach yields unbiased gradient estimates, it also suffers from …