Posts

Showing posts from November, 2019

Overfitting

Image
Modern research is all about building on top of existing knowledge. Given that there's such a high pile of pre-existing knowled ge, one really has to climb fast to get to the top of this pile, before one can start making original contributions. This could lead to one not learning some basic skills and knowledge: a super specialist in a narrow topic incapable of doing much else. While there's a reason why things at large have turned out this way, one must keep oneself aware of this, and try being hands-on on a day-to-day basis.