The exploration-exploitation dilemma is central in reinforcement learning: When facing a decision with uncertain outcome, should we try to learn more about the environment or act according to what we believe to be best? For my research, I am interested in understanding the theory behind exploration-exploitation trade-offs as it appears for example in Bayesian optimiztion, stochastic bandits and more general reinforcement learnig scenarios. In particular, I focus on developing partical algorithms with theoretical guarantees. Before joining the group of Andreas Krause, I have obtained a Master Degree in Mathematics at ETH Zurich.
