Neural networks is referred to a series of algorithms that try to mimic the neuronal structures of a human brain and thus try to identify underlying relationships in a set of data
Artificial neural networks are computational models based on the biological nervous system of the human brain. They can be defined as a set of processing units or computational components characterized as artificial neurons. Neural networks have the ability to acquire and maintain information based knowledge. The main function of neural networks is to receive a set of inputs which progressively perform complex calculations and solve a problem by its output. They aim to solve problems in a similar way as the human brain does. In contrast to the conventional approach of programming, neural networks learn from observational data and figure out their own answer for the problem.
In 2006 the discovery of techniques for learning in deep neural networks enabled a technological leap. Before, it was not possible to train neural networks and thus overcome conventional approaches. Today, this technology base is known as deep learning. By further development, deep neural networks became very promising and meanwhile show high performance.
Companies like Facebook, Microsoft and Google increasingly implement the technology in speech recognition and natural language processing to improve their products and services and make them more user-friendly.
There are several issues related to engineering and science where artificial neural networks can be employed: pattern recognition as applied in image, speech and writing recognition; prediction system which is about estimating future values such as stock market prediction; data clustering in which the goal is to identify similarities of given input patterns and allow its clustering.
Most likely all neural network technologies will become more sophisticated and considerably improved by researchers in the future. By developing better training methods and network architectures the future will allow common usage of self-driving cars, robots that can see, feel and predict things around them, self-diagnosis of medical problems and many other abilities.
Experts believe that neural networks and deep learning algorithms are still far from their theoretically possible performance but they also predict that the algorithms and technologies will become so efficient that they will be able to run on cheap mobile devices.
Literature: Artificial Neural Networks : A Practical Course / by Ivan Nunes da Silva, Danilo Hernane Spatti, Rogerio Andrade Flauzino, Luisa Helena Bartocci Liboni, Silas Franco dos Reis Alves