1. Research in system seems a bit different from what I thought about research. Instead
of starting with very comprehensive and sophisticated analysis, try to deliver some simple
model first and then add more complexity and intelligence into it incrementally.
2. Instead of constructing ideas in mind, try implementing it first, then make some
observation on the experiment results to obtain clues for some potential solutions.
3. Think from a broad perspective and in an adroit way. Do not get restricted to some
pre-defined rules or feedbacks. Nothing matters more as long as it works.
4. Always make sure the experiment results are repeatable.
5. Spend more time analyzing the experiment results. Do not repeat serial experiments
without any motivation.
6. Always start your work with motivation and keep in mind your focus and contribution.
And push your work towards your research goal effectively. Time is the most important, so do
not waste time on anything unrelated to your research goal.
7. If your objective gets complicated, try to decouple it, implement each sub-objective
one by one and integrate them into the system incrementally. Do not be hasty to expect to
finish the whole thing in one single run.
8. Discussion is necessary, peer feedback is usually useful. But do not eaily trust
anybody (even your advisor or senior partner). Think independently and critically based
on the experiment data, because only data tells you the truth.
9. If you have an idea, try to implement it (if in a few hours) to see if it works.
You lose only a few hours if it fails (usually you will gain some insights unexpectedly).
10. Be meticulous and suspicious. A small bug makes a huge difference to your findings.
11. Always make a short-term (i.e. weekly, daily) schedule and set some deadline (i.e. conference
deadline) to motivate yourself, even if your advisor has not yet set for you. Also, prepare a backup
plan for your current research.
12. You, not your advisor, manage your research project. So plan things (i.e. implementaion,
evaluation, paper writing) ahead by yourself before your advisor asks you to, work at high efficiency,
and get well prepared for the worse.
13. The merit of academic research does not lie in delivering a perfect solution, but some potentially
feasible way to fill up some niche.
14. Set your standard high and pursue research goals with energy, speed and focus.