Some years before a movie came out. It was a Spielberg’s one and the name was AI. Nice movie, nice story. But AI was not an unknown word even at this time when the movie came out. Artificial Inteligence was a known term and describes the concept of how a machine can learn and addapt from/to situations. It was first used from John McCarthy, in 1955 as “the science and engineering of making intelligent machines”.

But what lays behind this term? Neural Networks? Programmers in order to make a program “learn”  and find the best way to do something that it is designed to do, started copying human newral network structure. This is not very easy to be described but Jeff Heaton gives a very good example of how Neural Networks work in his youtube channel. If you cannot find the video this is the one someone should start: [youtube=http://www.youtube.com/watch?v=PNqc4fkdfIo&w=560&h=315]

In general, AI makes the use of neural networks in order to find patterns in the data that benn imported. With this process the program uses the patterns from previous situation in order to solve problems. A very good example (and easy for someone to understand it) is the Rock-Paper-Scissors game here. It is you against the computer. In the Novice mode computer learns from your kind of play. In the beggining its easy to beat it, but as the game goes on the computer makes a database with your playings. So when a new turn starts, the computer tries to find patterns from your previous playings and it is becoming more and more difficult fot the user to win. The veteran mode is difficult from the beggining as the computers searches from a huge database of other people’s playings for patterns that match your movements.

In general, Jeff Heaton explains how the Neural Networks work in a very easy way with not so many maths that would confuse someone. A great example of how Neural Networks work (in MatLAB) is this one below.

See, read and enjoy. Cheers!!

[youtube=http://www.youtube.com/watch?v=zEMZEpkemKg&w=420&h=315]