How to approach research in AI
How we advice our interns to go about research
This is what we shared with the research interns who joined Lossfunk recently. Crossposting it below if it helps others:
Reminder on research process.
I strongly recommend everyone to think and write
what research question are you exploring
a couple of research hypotheses accompanying them
Research hypothesis is your hunch about what’s true about research question.
Only once you have research hypotheses, you should think of experiments that will prove or disprove your hypothesis.
Without this, you’ll simply shoot in the dark randomly which is fine for initial exploration but eventually you need to methodically explore your research question
Also, when you’re running experiments make sure you know what’s the core experiment and what’s peripheral. Focus on core and make simplifying assumptions about every thing else.
If you don’t do that, you’ll simply conflate multiple factors and you’d never know why something works or doesn’t work.
- Paras is founder and researcher at Lossfunk.