Skip to main content

Read Research Paper

5 - 20 paper = basic understanding / good enough 50 - 100 paper = good understanding

Reading papers​

  1. Compile list of papers (medium/blog posts/arxiv)
  2. Skip around list
    • Read 10% of a list, throw away papers that don't make sense, read the good ones, read the papers referenced by the good papers

Reading one paper​

  1. Title / Abstract / Figures
  2. Intro / Conclusion / Figures / skim rest / (skip related work)
  3. Read but skip / skim math
  4. Read the whole thing but skip parts that don't make sense

Try to answer this questions:

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

Sources of papers​

  • Twitter
  • ML Subreddit
  • NIPS / ICML / ICLR
  • Friends

Undestanding Math​

  • Rederive from scratch
  • Download and run open-source code
  • Reimplement code from scratch

Tips​

  • Do steady reading not short burts

Source