When you hear the term neural networks, it gives you a sense that its a form of biological terminology pertaining to brains. And I have to tell you candidly that it's a no brainer to guess that and, in fact, we are treading along the right path by doing so. We will see how it is connected to that.
Neural networks have brought in a revolution in the data science world. Until 2011, due to not having enough computation power, the people rooting for neural networks were not able to propagate it to the extent that they wanted. But, with the advent of cheaper computation solutions and more research in the area of neural networks, they have taken the data science and artificial world by storm. Neural networks are an algorithm that can be applied in both supervised and unsupervised learning. With deeper networks, they are able to provide solutions to unstructured data, such as images and text.
In this chapter, we will cover the following topics:
- Neural networks
- Network initialization
- Overfitting
- Dropouts
- Stochastic gradient descent
- Recurrent neural networks