Those are the things from science that I want to concentrate on and systematically learn:

  • Reinforcement learning
  • Computational neuroscience

Both of these are closely related and even complete each other.

Both of these require solid math background, especially in statistics, and ODEs. I think that some elementary knowledge of electrodynamics wouldn’t hurt as well because computational neuroscience works with patterns obtained from neurons and their connections. Yes, a good knowledge of Fourier Analysis also wouldn’t hurt.

I think the good plan refreshing math background will be the following:

  • Go over ‘Introduction to Analysis’ of Maxwell Rosenlicht and refresh what I already know.
  • Learn Fourier analysis using ‘Fourier Analysis: An Introduction’ by Elias Stein (from
  • In parallel, study linear algebra using 'Linear Algebra Done Right’ by Sheldon Axler. This book is not necessary. Gilbert Strang’s book would also do I guess.
  • Watch 3blue1brown videos.
  • And practice, practice a lot. Solve a lot of problems. Perform analysis on some ‘real’ data.

to be continued…