Bits of string: academic bits

If you’re doing a master’s degree with me please see this note: (202508281445). If you’re doing a master’s degree with someone else probably still see that note.

Writing papers and doing research

(§) The Kajsa Møllersen u-net architecture for writing papers or paper unet for short. Certain sections are connected across the paper (“skip connections” in the lingo) like how basically the conclusion reflects the abstract (“we did so and so and found so and so, etc., etc.”).

(§) In paper unet methods and results are at the bottom and really they are the basis on which a paper is built. See the Jeff leek guide to writing papers where the whole thing starts with doing the research and making a handful of solid figures to write around. That is to say get the results and write from the inside out (or the bottom up in paper unet). Your abstract+intro are just speculation before the methods+results are in place.

(§) How to ML paper is a good guide to writing a methods type paper as well.

(§) A modest proposal: I think every research group in the department (of computer science in my case) should be forced to write at least one paper with every other research group in the dep’t. Not only will this produce $\Omega(n^2)$ papers but it will either unite the previously insular research groups throught collaboration and mutual respect or it will unite the previously insular research groups in shared animosity toward whoever decided this was a good idea.

Teaching

(§) NTNU has a conversion from a 100 point scale to letter grade as follows:

A: 89–100 
B: 77–88
C: 65–76
D: 53–64
E: 41–52
F: 0–40

Backlinks:

• (202508181349) Bits of string