I am interested in exploring how agents can learn to abstract structure out of tasks and use this to adapt better to different classes of tasks. Thus, I am currently working on gaining further insight into generalization and domain adaptation in Reinforcement Learning and Meta-Learning. My greater research ambition is to develop agents that can learn how to tune their own algorithms. My current research interests are:

• Contextual Reinforcement Learning
• Automated Reinforcement Learning
• Meta Reinforcement Learning