Analyzing the Effects of Reinforcement Learning to Develop Humanoid Robots

Abstract

Reinforcement learning is a flourishing machine learning concept that has greatly influenced how robots are designed and taught to solve problems without human intervention. Robotics is not an alien discipline anymore, and we have several great innovations in this field that promise to impact lives for the better. However, humanoid robots are still a baffling concept for scientists, although we have managed to develop a few great inventions which look, talk, work, and behave very similarly to humans. But, can these machines actually exhibit the cognitive abilities of judgment, problem-solving, and perception as well as humans? In this article, the authors analyzed the probable impact and aspects of robots and their potential to behave like humans in every possible way through reinforcement learning techniques. The paper also discusses the gap between 'natural' and 'artificial' knowledge.