Why Machine Learning?
In the last few years, mankind has faced a very serious problem: the big data problem. In only 10 minutes today we produce information as much as in the last 100 years. The consequence is clear: we arrived at the point when the human brain is no more able to process the huge amount of available information, and hence the need of help from machines equipped with proper learning software, which are able to digest the information in a timely manner.
Why Deep Learning?
The successful tract so far is lead by the deep learning, which mainly started in 2012, and had emerged from an older trend, called Artificial Intelligence. This comes with a change in the paradigm of how the computer programs are constructed. In a classical program, all rules are set by the programmer, while in the case of neural networks, the computer has the possibility of learning from provided data. The computer is trained to learn, instead of being previously programmed to apply certain fixed rules. This is done with the help of neural networks.
What is a NN?
A neural network is a collection of computing units, or neurons, which are connected together, in a more, or less similar way as the neurons in the human brain. This structure allows neurons to send messages among themselves and consequently, to straighten those connections that lead to success in solving a problem and diminishing those which are leading to failure.