In 2-and-a-bit years I wrote 12 papers, received 4139 GitHub stars, got 3271 Twitter followers, authored 1 textbook – doing double-duty as my th

Just know stuff. (Or, how to achieve success in a machine learning PhD.)

submited by
Style Pass
2023-01-27 16:00:08

In 2-and-a-bit years I wrote 12 papers, received 4139 GitHub stars, got 3271 Twitter followers, authored 1 textbook – doing double-duty as my thesis – and got the coveted big-tech job-offer.

If you’re interested in a textbook on Neural Differential Equations with a smattering of scientific computing, then my thesis is available online.

Quite a few folks seem to have looked at this, and messaged me – mostly on Twitter or Mastodon – asking for advice on how to achieve success in a machine learning PhD?

Now, I don’t think “Just Know Stuff” is a terribly controversial opinion – undergraduate classes are largely based around imparting knowledge; the first year of a new PhD’s life is usually spent reading up on the literature – but from the number of questions I get it would seem that this is something worth restating.

Know your field inside-out. Know as much about adjacent fields (in math, statistics, …) as you can. Don’t just know how to program; know how to do software development. Know the mathematical underpinnings your work is built upon. And so on and so on. Indeed: possessing a technical depth of knowledge is how you come up with new ideas and learn to recognise bad ones.

Leave a Comment