Intelligent Data Analysis

5Intelligent Data Analysis: Deep Learning and Visualization

Than D. L¹ and Huy V. Pham²

¹University of Bordeaux, Labri, Bordeaux, France

²Ton Duc Thang University, Faculty of Information Technology, Ho Chi Minh, Vietnam

5.1 Introduction

Deep learning [1] and deep reinforcement learning [2] are currently the resolution in applied artificial intelligence for many applications based on upgrading excellent performances. Basically, deep learning [1] is the best model for data representation by understanding and reinforcement learning [2] is a modern approach to solving the decision making. These are essential to represent the basic things forming the intelligent autonomous systems [3] that can be enabled to solve the basic level based on simultaneous localization and mapping (SLAM), which is based on interacting with the unknown environment. Deep learning is sometimes called hierarchical learning or deep structured learning. The history of deep learning and neural networks is not new. Indeed, the first mathematical model of neural networks was introduced by Walter Pitts and Warren McCulloch in 1943 [4]. However, it grew up just a few years ago by upgrading to graphic processing units (GPUs), which are increasing more opportunities for many applications in artificial intelligence. Figure 5.1 shows us the wide range of deep learning.

Let's look at the progression of fields of deep learning. In machine learning, there commonly exist three types of learning: supervised learning, unsupervised ...