Basic Understanding of Artificial Neural Networks

By POOJA BISHT |Email | Mar 8, 2019 | 12165 Views

Have you ever noticed how particular information reaches your brain and how did it happen to take a decision or react thereafter? How do you feel something is hotter or colder after touching it? How does your brain recognize a sound? What it all takes to think? How do you perform reflex action? 

However, we think it just a normal as anything we do daily, there are many complex reactions and processes that are actually doing a lot to produce these reactions, decisions, and feelings. 

If you haven't still guessed about what must I be talking about or never heard about these processes then I must make you familiar about those important nerve cells that actually constitutes the brain, the Neurons. Neurons are an integral part of the brain which helps us to see, hear, smell & feel (there are several other functions as well). They are like the chemical messengers which actually help in transmitting signals to the brain to make it identify a particular situation. There are millions of neurons that constitute a brain.

The concept of Artificial neural network comes after researching on these biological neurons only, identifying the complex processes so to develop a similar like that for AI machines. Neural networks are nothing but a system of neurons contained in it. Researchers and great minds were always amazed by the fact how the human brain works and tried to develop a neural network which works the same as that of biological neuron and also got some success in that. The aim behind this idea was to solve the complex problems that machines needed to solve and that too unsupervised. 

It was Warren McCulloch and Walter Pitts in 1943 who created a computational model for neural networks which was based on mathematics and algorithms. This work led to work on nerve networks.

An artificial neural network is a network which is designed to solve the complex problems, unlike the simple problems where the input is fed and output generated which is supervised by humans. It has got the inspiration from the biological neural network as has been discussed and contain artificial neurons in it. These networks generally provide output by considering the information or examples of the input provided to it. Like the image recognition where you would provide the system some information about the particular thing (it could be its name as well) that you want the system to identify and after feeding some examples of it into the system and the system finally identify any image particular to that thing. 

Artificial neural networks are being used for a variety of purposes like image recognition, computer vision, speech recognition etc.

An ANN being which is used in the character recognition could recognize the handwriting which has so many applications in contemporary time. Similarly, for Image recognition, an ANN is doing better to recognize images and found its applications in the industries where such software are in demand.

The importance of using an Artificial neural network is that it produces the output of the data by learning from the inputs provided to it, so we don't have to feed the input again and again. Also, it provides solutions to the complex problems which are needed to solve in the contemporary time.

Source: HOB