Department of Image Processing

The seminar will focus on deep reinforcement learning methods for the utility optimization problem. In contrast to the standard case when agent maximizes expected utility, the seminar will focus on agents maximizing a quantile of the utility stream. The seminar will illustrate the methods on a simple consumption-investment decision problem. As a part of the seminar, we will review the reinforcement learning, it’s extension to deep reinforcement learning and introduce the quantile preferences.