A Comprehensive Study on Reinforcement Learning and Deep Reinforcement Learning Schemes

Abstract

Reinforcement learning (RL) has emerged as a powerful tool for creating artificial intelligence systems (AIS) and solving problems which require sequential decision-making. Reinforcement learning has achieved some impressive achievements in recent years, surpassing humans in a variety of areas. According to recent research, deep learning (DL) techniques are used with techniques of reinforcement learning to recognize meaningful identification for a problem regarding high dimensional raw data input & enough to solve artificial general intelligence (AGI). In addition to the main concepts, this paper highlights the intuition behind the use RL and deep Q-network (DQN) over other algorithms. In this research paper, different methods and details for dealing with reinforcement learning difficulties have been presented. Finally, various difficulties of the reinforcement learning have been addressed.

Keywords

• Reinforcement Learning (RL)
• Artificial Intelligence Systems (AIS)
• Deep Q-Network (DQN)
• Artificial General Intelligence (AGI)
• Deep Reinforcement Learning (DRL)