What I Thought About Data and the Physical Sciences and Why I Was Wrong

I thought there was no way that you were going to be able to apply neural networks, in particular, or machine learning techniques in general, to understand physical systems.