Nadav Merlis

Latest

• On Bits and Bandits: Quantifying the Regret-Information Trade-off
• Never Worse, Mostly Better: Stable Policy Improvement in Deep Reinforcement Learning
• Reinforcement Learning with a Terminator