I am interested in how our brains learn by relating new concepts to those we’ve previously learned. In machine learning, this is studied under the area of inductive transfer or transfer learning. I’m particularly interested in transfer learning within the context of the reinforcement learning paradigm. Different than learning to classify an input, a reinforcement learning agent receives observations and rewards through interactions with an environment. Given an assumption of rationality, the agent acts in that environment in order to maximize its long-term gains. Learning in this paradigm amounts to determining the appropriate policy for determining what action to take given an observation about the state of the environment. The combination of the structure and dynamics of the environment with the rewards structure define the task.