A Review on Deep Learning Models

Abstract

Deep Reinforcement learning a subpart of deep learning is used to maximize a numerical parameter expressing long term objective to control a system. The main aim of reinforcement learning also known as semi-supervised learning is to develop learning algorithms which are efficient as well as learn about the limitations of this algorithm. In certain networks like Internet of Things and Unmanned Aerial Vehicle, decisions to maximize the network performance is expected and deep reinforcement learning has been effectively used in achieving optimal policy like actions and states. Reinforcement learning has vast applications in the field of artificial intelligence, control systems, robotics, genetics, statistics, etc. The combination of this algorithm along with the neural networks can solve complex problems. The combination of deep learning along with the reinforcement leaning is used to overcome the complex and large scale network problems like wireless caching, network security, data rate control etc. this paper discusses about the reinforcement learning and its various algorithms along with its applications in various industries.