I wanted to discuss methods for taking notes, since they’re an important tool in ML.
Most of us take notes when working through a paper, or a concept. Then we make use of these notes when writing our own papers or software.
I feel that notes are most effective when the medium you use to write them is expressive (so you can write down equations / diagrams easily), as well as searchable, so you can review your own notes easily. What note taking methods have been useful for you? Feel free to go into as much detail and reasoning as you’d like.
Personally I haven’t found a solution that is ideal. I find that I never search through my physical notebooks, and I feel limited and slowed down trying to write in Jupyter notebooks.