Ever since the beginning of the computer era, humans have been trying to mimic the brain into the machine. Researchers have been developing methods that would make machines not only compute but also decide like we humans do. This quest of ours gave birth to artificial intelligence around the 1960s. By definition, artificial intelligence means developing systems that are capable of accomplishing tasks without a human explicitly programming every decision. In 1956, the first program for playing checkers was written by Arthur Samuel. Since then, researchers tried to mimic human intelligence by defining sets of handwritten rules that didn't involve any learning. Artificial intelligence programs, which played games such as chess, were nothing but sets of manually defined moves and strategies. In 1959, Arthur Samuel coined the term machine...

Though machine learning has provided computers with the capability to learn decision boundaries, it misses out on the robustness of doing so. Machine learning models have to be very specifically designed for every particular application. People spent hours deciding what features to select for optimal learning. As the data cross folded and non-linearity in data increased, machine learning models struggled to produce accurate results. Scientists soon realized that a much more powerful tool was required to apex this growth. In the 1980s, the concept of ANN was reborn, and with faster computing capabilities, deeper versions of ANN were developed, providing us with the powerful tool we were looking for—deep learning!

This section is where you will begin the journey of being a deep learning architect. Deep learning stands on the pillar of ANNs. Our first step should be to understand how they work. In this section, we describe the biological inspiration behind the artificial neuron and the mathematical model to create an ANN. We have tried keeping the mathematics to a minimum and focused more on concepts. However, we assume you are familiar with basic algebra and calculus.

As we mentioned earlier, deep learning is inspired by the human brain. This seems a good idea indeed. To develop the intelligence of the brain inside a machine, you need the machine to mimic the brain! Now, if you...

Before proceeding any further, let us quickly set up our coding environment. This book uses Python programming language all throughout the chapters. So, we expect you to have prior knowledge of Python. We will be using two of the most popular deep learning open source frameworks—TensorFlow and Keras. Let's begin with setting up Python first (in case you don't have it installed already).

Let's take a quick look at what we learned in this chapter. We began by briefly discussing artificial intelligence and its evolution through machine learning and then deep learning. We then saw details about some interesting applications of deep learning like machine translation, chatbots, and optical character recognition. This being the first chapter of the book, we focus on learning the fundamentals for deep learning.

We learned how ANN works with the help of some mathematics. Also, we saw different types of activation functions used in ANN and deep learning. Finally, we moved to set our coding environment with TensorFlow and Keras for building deep learning models.

In the next chapter, we will see how neural networks evolved into deep feedforward networks and deep learning. We will also code our first deep learning model with TensorFlow and Keras!