Intrinsic Motivation in Model-Based Reinforcement Learning: A Brief Review

The reinforcement learning approach offers a wide range of methods to the solution of problems of the control of intelligent agents. However, the problem of training an agent from sparse rewards remains relevant. One possible solution is to use methods of intrinsic motivation, an idea that comes from developmental psychology. Intrinsic motivation explains human behavior in the absence of extrinsic control stimulate. In this article, we reviewed the existing methods of determining intrinsic motivation based on the learned world model. The systematization of modern works in this field of study was proposed. This system consists of three classes of methods differing according to the application of the word model to the agent components of reward system, exploration policy, and intrinsic goals. We proposed a unified framework for describing the architecture of an agent using a world model and intrinsic motivation to improve learning. The prospects for development in this field of study were analyzed.