Applied ML Research checklist

Share on:

I keep coming back to this great piece by John “OpenAi” Schulman : An Opinionated Guide to ML Research.

Here I want to quickly checklist his advice and whether I follow it or not:

  1. Choosing problems
    • Honing your taste
      • Optimise your environment to get input from other researchers
      • Reflect on what research is useful and will pass the test of time
    • Idea-driven vs Goal-driven research
      • Formulate a (per-project) goal that drives you
  2. Making continual progress
    • Notebooks
      • Keep a daily notebook
      • Review said notebook every fortnight and every couple of months
    • Switch problems in a controlled way
      • Allow 1 day per week on something totally different from your main project (~ epsilon greedy for Multi-arm Bandits ; ~ 20% time project)
  3. Personal Development
    • build your knowledge of ML
      • Read textbooks and papers
      • Reimplement algorithms from these sources

I do read textbooks and reimplement algos from textbooks and papers. I also hold manuscript and digital research journals ever since my internship advisor at Bosch in Singapore taught me so in 2014.

2 areas I want to improve are: going back to practising 20% time project (although I do that on weekends basically) and get more input from other researchers.

“If you’re not fortunate enough to be in an environment with high density of relevant expertise, don’t despair. You’ll just have to work extra-hard to get ahead of the pack, and it’s extra-important to specialize and develop your own unique perspective.” - John “OpenAi” Schulman