Deep learning is an advanced segment of machine learning, that focuses on neural networks. Inspired by the human brain, the neural network teaches computers how to analyze data through experiences. It uses neurons or networked nodes arranged in a layered framework to mimic the structure of the human brain.

Learning from experiences: Instead of providing step-by-step instructions to the computer, we input numerous examples and allow the computer to analyze these examples and identify patterns independently. 

Hierarchy of Concepts: Think of learning in layers. The computer starts by understanding very simple ideas. It then uses these simple ideas to understand more complex ones. For example, to recognize a face, it might first learn to see lines, then shapes, then parts of the face, and finally the whole face.

We do not have to program every detail. The computer learns by itself from the data we provide, much like a child learns from exploring the world. It can identify intricate patterns in images, text, sounds, and other types of data to generate precise insights and estimations, There are three types of Deep Learning Models; Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM)

ML v/s DL: Machine learning needs data to be well-organized and labelled, while deep learning can handle messy data like images, and learning patterns on its own without much human input.