Self-supervised learning model

This paper describes a reinforcement learning algorithm based on supervised learning. Many of reinforcement algorithms use associative search to discover and learn actions that make the system perform a desired task. One problem with associative search is that the system's actions are often inconsistent. In the searching process, the system's actions are always decided stochastically. Therefore, the system cannot perform learned actions more than once. To solve this problem, this algorithm uses a neural network which can predict an evaluation of an action and control the influence of the stochastic element. The effectiveness of this algorithm was checked by computer simulations