“Science is a beautiful gift to humanity; we should not distort it.”
(A. P. J. Abdul Kalam)
Machine learning copes with data. All scenarios: supervised, semi-supervised, unsupervised and reinforcement-based extracts pieces of information from the data and tries to generalize the rules they learn when unseen samples are presented. Even if it can be either a limitation or something absolutely obvious, I’ve formalized a concept I consider as a pillar of any Artificial Intelligence research:
A natural or artificial intelligent agent can learn only what is already encoded in the data it perceives.
Another statement that I consider almost essential for Machine Learning is:
If a programming language needs more than one line of code to show “Hello World”, it will be probably a bad choice for Machine Learning.