Skip to content

Instantly share code, notes, and snippets.

Last active November 30, 2020 01:11
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
Star You must be signed in to star a gist
What would you like to do?
[Stanford CS230: Deep Learning | Autumn 2018 | Lecture 8 | Career Advice / Reading Research Papers]#stanford #cs230 #deep_learning #research

Reading research papers

  • Compile list of papers. (arXiv, medium/blog posts.)
  • Skip around the list.
    • 5 - 20 papers: a basic understanding of an area.
    • 50 - 100 papers: a very good understanding of an area.

Reading 1 paper

Take multiple passes through the paper:

  1. Read the tittle / abstarct / figures (there are a lot of research papers where the entire papers is summarized in one or two figures in the figure caption);
    • Just by reading the title, abstract and the key neurual network architecture figure that just describes what the whole papers are, and maybe one or two of the experiments section.
  2. Intro + Conclusions + Figures + Skim rest (skim related work);
  3. Read but skip/skim math;
  4. Whole thing but skip parts that don't make sense.

When you've read and understood the paper, theses are questions to try to keep in mind:

  • What did the authors try to accomplish?
  • What were the key elements of the approach?
  • What can you use yourself?
  • What other references do you want to follow?

One of the principles is:

Go from the very efficient high information content first, and then go to the harder materail later.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment