It all starts with a single neuron! Or with a perceptron. Perceptron is a neuron’s computational model. If you link some of them together, you get to a real ‘artificial brain’ capable of learning complex stuff.
When I try to explain basic AI concepts to non computer scientist, I usually start like this …
You have a baby that knows nothing. At the beginning, the baby makes many mistakes, but improves over time until sufficient grown. This iterative process of trial and error, in human terms, is called growing. In Machine Learning it is called learning.
A while ago, I had some spare time and wanted to actually see the learning process of a machine. So I made a small visualization of perceptron learning – with Python implementation. Just as a remainder, the Neural/Deep Network/Learning starts right with a single perceptron, so this is a good entry point into the world of Machine Learning and AI. Enjoy!
Full source can be found here.
Perceptron Learning in Python – Neural Networks start here. by Dženan Hamzić
Bits & Notes 