Abstract

In recent years, deep reinforcement learning has been developed as one of the basic techniques in machine learning and successfully applied to a wide range of computer vision tasks (showing state-of-the-art performance). In this tutorial, we will overview the trend of deep reinforcement learning techniques and discuss how they are employed to boost the performance of various computer vision tasks (solve various problems in computer vision). First, we briefly introduce the basic concept of deep reinforcement learning, and show the key challenges in different computer vision tasks. Second, we introduce some deep reinforcement learning techniques and their varieties for computer vision tasks: policy learning, attention-aware learning, non-differentiable optimization and multi-agent learning. Third, we present several applications of deep reinforcement learning in different fields of computer vision. Lastly, we will discuss some open problems in deep reinforcement learning to show how to further develop more advanced algorithms for computer vision in the future.

Keywords:Deep reinforcement learning, policy estimation, non-differentiable optimization, weakly supervised learning.

Time:This tutorial will be presented on the afternoon of June 17, 2019.