What Is Deep Learning?

Deep learning is a subset of Machine Learning AI technology. It emulates the human brain in order to process data, extract patterns and make decisions.

Human Neuron Operation

A human brain contains neurons shown schematically in the diagram below. Dendrites receive signals from other neurons in a brain. The cell body combines the received signals. It generates an electric pulse if the sum exceeds a certain level. The resulting signal travels down the Axon to Axon tips. These in turn distribute this signal to other neurons by releasing neurotransmitters such as Dopamine, Glutamate, Serotonin, etc… There are approximately 40 neurotransmitter types. The released neurotransmitters are picked up by dendrites of multiple other neurons. Each of those combine the signals they receive and process them and send signals to other neurons. This goes on and on, and eventually causes a release of hormones or other actions in a body. 

Components of a neuron
Key components of a neuron

The network of neuron is shown schematically in the diagram below. In this diagram, electrical signals from an eye feed into the dendrites of the brain visual cortex (he first column of dendrites in the diagram). Each neuron processes the signals it receives and distributes them to other neurons in the brain.

Neuron connections
Brain-eye connections

A human brain contains approximately 86Billion neurons each connected with up to 10,000 other neurons creating a network with hundreds of trillions of connections.

Deep Learning Neural Network

A Deep Learning system is similar to a network of neurons in a brain. Logical nodes which are software routines perform a function similar to neuron cell body.  This is called a Neural Network.

Convolutional Neural Network
Convolutional Neural Network

This diagram shows a Neural Network in the context of image analysis which is similar to how a brain deals with signals from an eye. The Input Nodes (the first column of circles) are like the Dendrites of the human visual cortex neurons. Deep Learning neural networks employ an intermediate layer called Convolution Layer shown here as box. They are called Convolution Neural Networks. Convolutional Networks are pretty much what all Deep Learning systems employ today.

Convolutional Neural Networks

The box called the Convolution Layer represents a complex arrangements of software algorithms which perform a role of image filters. This too is analogous to how a human brain works. Certain neurons fire when an eye is looking at a vertical line. Other neurons fire when an eye sees a horizontal line. This means that the visual cortex neurons specialize in recognizing different shapes. Deep Learning scientists have emulated this functionality. They created logical mini-algorithms which function as filters. These filters detect various image parameters such as lines, shapes, textures, edges, colors, etc….  There are over 100 different filter designs. Convolutional neural networks use combinations of filters depending on application. We will discuss this in more detail in subsequent tutorials.

Although Deep Learning Convolution Neural Networks have much fewer nodes than a human brain they can be quite complex. It is common to find networks with a million or more connections. Even though the number of connections is much smaller than in a brain, Deep Learning networks can perform amazing tasks. It can recognize images, sounds, understand languages and speak without human involvement.

Deep Learning Networks Require Training

There are two categories of Deep Learning systems Supervised and Unsupervised. Supervised learning require training the system. For example, labeled images of cats, dogs, tumors, clouds, cars, trees, etc… are used to train a neural network so it can recognize such images later on. This is similar to how parents teach their children. Unsupervised networks do not require training. They take unlabeled data and make discoveries by trial and error much like humans and animals do. We describe both types of networks in other tutorials.  

Read more …