I am a (soon-to-become, to be honest) theoretical physicist. I want to learn a bit about AI. So as you know in physics we develop theories based on as few and as simple basic equations as possible which shall explain as much of the experimental results and observations as possible. I feel that this is kind of not how AI solves problems.
My understanding is that AI can be understood as a very generalized and abstract statistics software package handling input data in a general way to find the "best fit" to some form of problem. Is that correct? I know it isn't. But is it vaguely correct?
I give you an example. In weather prediction there is a technique called MOS (model output statistics). It collects output from numerical weather prediction models (simulation software) as well as observational data and finds statistical relations between them to correct the model output for errors. For example, it might be that the intensity of precipitation in London is on average underestimated by the model by 10 %, so MOS will correct for that. Over time, it improves itself, because it collects more and more data. Is this already a form of AI?