Towards Efficient Transfer of Tacit Strategies for Asset Management: A Sequential Decision-Making Perspective

To achieve our aim, in this project, we will integrate skill learning and sequential decision-making under uncertainty. We will consider the following simplified problem. Suppose a robot is given a task to accomplish and a set of trajectory data, where the trajectories are execution traces of the robot performing an optimal policy to accomplish a variety of simpler tasks that are related to the given task.