My first semester of MTech at IISc got over on 29th November. I thought of writing a post on some learnings of first semester during exam time as exams usually make us realise what did we do wrong or what we did right during whole semester. Apart from learning that most of the research life is just about ‘bounding norms’, I got to learn some important things that I was never aware of before starting my master’s. Here they go:

  1. ‘Optimum myth’: There’s no such thing as ‘optimum routine’ that can help you do things perfectly. You don’t have to plan and write thousands of time what your routine should look like. Things become optimum when you just focus on tasks with priorities and just do them on that basis. One needs to do more instead of spending whole day learning something. It’s like online learning: Making mistakes, learning, committing at the same time, resulting in some regret bound at the end of the day with optimal policy at the end. That policy may change next day but that is just because of uncertainties one didn’t take into account previously. Sometimes some part of that online learning may need to be offline but it should be balanced according to one’s priorities.

  2. ‘Perfection myth’: In order to learn efficiently, you need to first start with the problem and then see what is it that you really need to learn to tackle this problem rather than first learning all techniques and then approaching the problem. The latter approach will make you lazy and unmotivated while first one will help you learn things quickly and make you feel that you are doing something useful. It works for me atleast. For example: While bounding regret for UCB algorithm, you don’t have to know all kinds of concentration inequalities. You should first get an idea of what is it that might be useful for you to learn in order to approach the proof efficiently. That way, you won’t sacrifice on your intuitive thinking and at the same time will learn to use Maths in proper way.

  3. ‘Finish fast myth’: I always had this tendency (still have) to finish things fast. I said to myself in the beginning that I will complete all courses in first month itself but when I couldn’t, I would feel that I’m some loser. The point is: Coursework is not just to let you know things but rather to let you become comfortable in thinking about things without losing patience. It should be a regular thing to solve problems, learn new things, make projects, etc.

  4. The last one: If you’re in research program, you need to balance between coursework and your research progress. So make sure to take out atleast one hour for research each day. It will help you to not lose complete touch with your work and also will act as a regular motivation in your coursework. Research degrees are harder in a way that you get only first year to complete all required credits. So make sure you maintain good GPA since beginning itself. That’s the only time you get to do so anyway.

I hope anyone who is looking for advise on managaing things in IISc finds this useful.

Happy learning!

PS: Yes, the first post on optimization and Bandit algorithms will come soon. Stay tuned!

Update on 20-12-2024

I forgot to add the most important point that has always helped me (Idk how general this advice is): Don’t copy anyone else’s method of learning or doing things. Always figure out your own style, own pace, and own goals. Not everyone takes a course or works on some problem for one single goal. Everyone has different intentions. Figuring out what you want from anything and how is it helping in your research progress is something that will always help you. Don’t stress yourself with too much courseload. Taking time to figure out what ‘optimal’ is always does wonders.