How to improve the machine learning understanding?

I mean the understanding of the theory part especially the learning theory.

Some people mentioned that if it is between the 1970s to 1990s, we can read Vapnik’s “Statistical Learning Theory”. If the 2000s, we can read Mehryar Mohri’s “Foundations of Machine Learning”. Theory after the year of 2010, we can read ICML/NIPS/ICLR/COLT.

I didn’t do that. However, I do find that there are some good papers in ICLR. Machine learning is such a board area and it is developing so fast. We will see what will help. What I want to say is reading the original paper instead of some summary articles will help a lot in better understanding this area.

ML is a hot area. Using a Classical Chinese writing。

A boat sailing against the current must forge ahead or it will be driven back.

学如逆水行舟，不进则退。

Cheers.