I understand that there is a lot of misunderstanding about artificial intelligence, so I would like to explain some basic things.
The foundations of Artificial Intelligence (a.i.) are two software areas: Expert Systems and Neural Networks. Expert Systems have a small niche and Neural Networks have become the core of almost any a.i.
For a basic understanding of how a neural network works, please read here.
What I want to point out here is that we are talking about software, anyway and software is a process that works through patterns. And patterns are a form of representation of reality. A pattern recognises repeated design problems which recur but always in a slightly different way.
When you realise that all your design problems can be matched to a pattern and that these patterns interlink with each other, then you have a kind of language that guides you through the process you are managing (diagnosis, catalogue actions, design process, text and image recognition, etc. are all pattern recognition. And the stock market forecast as well!).
Neural networks learn what you teach them, so it’s active the pattern “garbage in, garbage out”. You can teach manually to a neural network or through a software process, where a set of detection rules is active. r.Virgeel is trained with both techniques, i.e.
That’s it. Once you get the pattern concept and apply it using brute force, you can extract from the data a surprisingly precise response. It is simply the fact that our human brain cannot manage the quantity of data and relationships required to get the response, even if we can manage the process to produce the response.
Pattern recognition is spreading, as it opens a totally new field of applications.