Understanding Deep Learning and Neural Networks

Introduction to Deep Learning:

Deep Learning is getting lots of attention these days. Its making a big impact in areas such as computer vision and natural language processing. Lets us first try to understand the basics of Deep learning. Deep Learning is a machine learning technique that learns features and tasks directly from data. Data can be images, text or sound. Deep learning is often referred to as end to end learning.

For example, let’s say we have a set of vehicle images and we want to categorize which set of category does each image belongs to: cars, trucks, or bicycles, etc. We start with a labelled set of data (also called training data). The labels corresponds to the desired output of the task. The deep learning algorithm needs these labels as they tell the algorithm about the specific features and objects in the image. The deep learning algorithm then learns how to classify input images into the desired categories. We used the term “end to end learning” because the task is learned directly from data.

This diagram below explain this concept in a much clear manner.

Sample Use Case: Deep Learning to recognize zip codes:

Many of the techniques used in Deep Learning today have been around for many decades. For example, deep learning has been used to recognize handwritten postal codes in the mail service since the 1990s.

Factors affecting rise of deep learning:

The use of deep learning has increased hugely in the last few years due to various factors.
Accuracy : Deep Learning methods are now more accurate than people at classifying images.
GPU: GPUs enable us to now train deep networks in less time, and
Dataset: Large amounts of labelled data required for deep learning has become accessible over the last few years.

Neural Networks:

Artificial Neural Networks (ANN):

The term neural has been derived from human nervous system’s basic functional unit called “neuron”. A neural network is basically a highly interconnected network of billions of neurons with trillions of interconnections between them.

Neural networks are remarkably efficient tools to solve a number of really difficult complex problems. The first applications of neural networks usually revolve around classification problems. Classification means that we have an image as an input and the output is let’s say a simple decision whether it depicts a cat or a dog. The input will have as many nodes as there are pixels in the input image and the output will have two units. We will look at one of these two that fires up the most to decide whether it thinks it is a dog or a cat. Between these two there are hidden layers where the neural network is asked to build an inner representation of the problem that is efficient at recognizing these animals.

Types of Neural Networks Architectures :

Deep Learning uses Neural Networks

Most deep learning methods use neural network architectures. This is why we often hear deep learning models referred to as deep neural networks.

One popular type of deep neural network is known as a Convolutional Neural Network (CNN).

Convolutional Neural Network(CNN)

A CNN is a specially well suited for working with image data. The term deep, usually refers to the number of hidden layers in the neural network. Traditional neural networks only contained two or three hidden layers , some of the recent deep networks have as many as one hundred and fifty layers.