What is Deep Learning?
Deep learning is a subset of machine learning and is inspired by the human brain. Deep learning is based on artificial neural networks.
A typical deep learning model has at least three layers. Just like how humans learn from experience, the deep learning models also learn from each iteration and tweak the parameters accordingly to improve the outcome. Each layer learns from the previous layer and then passes its output to the next layer.
A Neuron is the basic unit of computation in a neural network. It is also called a node or unit. The neurons are grouped into three different types of layers:
- Input Layer
- Hidden Layer(s)
- Output Layer
The first (leftmost)layer in this network is called the input layer, and the neurons within the layer are called input neurons. It is used to provide the input data or features to the network. Input data can be in the form of text, image, or sound. For an image, the input might be a matrix of pixels and a typical audio signal can be expressed as a function of Amplitude and Time.
The middle layer(s) is called a hidden layer(s) since the neurons in this layer are neither inputs nor outputs. The number of hidden layers is termed as the depth of the neural network and deeper networks can learn more complex functions. The word ‘deep’ in deep learning refers to the number of layers through which the data has to go through before the output layer.
The last (rightmost) output layer contains the output neurons. This is the layer that gives out the predictions. The activation function to be used in this layer is different for different problems.
Unlike machine learning algorithms, the performance of deep learning neural networks continues to increase as we construct larger neural networks and train them with more data. In machine learning techniques performance reach a plateau after a particular amount of data.
One of the differences between machine learning and deep learning model is the feature extraction. In machine learning, feature extraction is done by humans whereas a deep learning model figures it out by itself without human intervention.
Deep learning architectures have been applied to social network filtering, customer support, care, self-driving cars, image recognition, financial fraud detection, speech recognition, computer vision, medical image processing, natural language processing, and many other fields. In short, our lives are influenced by deep learning on a daily basis because deep learning models are everywhere!
Happy Learning!
