Deep Learning Algorithm and Its Applications to IoT and Computer Vision

Abstract

Deep learning, which is the next phase of machine learning in recent decades, has significantly changed the way in which computer systems interpret human-centric content such as images, video, speech and audio. Different models have been introduced based on learning techniques such as supervised, unsupervised, reinforcement, and it is expected to accelerate and create even more innovative models in the coming years. With the rise of the Internet of Things (IoT), many real-time applications collect data about people and their environment using IoT sensors and feed them into deep learning models to enhance the intelligence and the capabilities of an application, for offering better recommendations and service. Experimental results of most of these applications looks promising when compared with traditional machine learning approaches. So, the main objective of this chapter is to make a self-contained review of Deep Learning (DL) models, starting with the Convolutioanl Neural Network (CNN), Recurrent Neural Network (RNN), Long-Term Short Memory (LSTM), AutoEncoder (AE), Generative Adversarial Network (GAN) and Deep Reinforcement Learning (DRL). Additionally, for providing better understanding of the models and their efficiency, we have added recently developed DL tools or framework and their applications.