Must Read: How to pursue research by Prof. Chandra R Murthy, IISc

23 AUGUST 2008

Hi All,

Its my great pleasure to have Chandra R Murthy, Assistant Professor at ECE department, Indian Institute of Science, Bangalore to talk about “How to pursue research”. This is one of the best interviews at Coffee With Sundar undoubtedly. Professor has a been kind enough to spare a lot of time in answering these questions. I hope his responses help you clear myths about pursue research! Without further ado, its Coffee With Chandra R Murthy.

Coffee With Chandra R Murthy

Me: Sundar Rajan G S CM: Chandra R Murthy

• Me: A student has decided to do research – how do u suggest that he does a self analysis to really find out whether he is really interested? (peer pressure is very common in colleges, so he might have resorted to research just because his friends are into it)
• CM: This is an excellent first question! Perhaps, the best way for me to answer it is to talk a bit about what I think it takes to do research. Then, the student can determine on his or her own, whether or not research is their cup of tea.

Research is the process of discovering or inventing something new, with the keyword being new. How would you know that something is new? By exhaustively surveying and studying what is “old”, of course. So, the first step in any research is to look up google, IEEE Xplore, CiteSeer, etc (and any other search engine you might fancy) and dig up anything that was ever published in your field and make sure you know about it. (Of course, I am assuming you have already picked the topic you want to do research in, but more on that later.) You may not have to read every one of them at the same level of detail, but certainly you do need to know the gist of what every paper says. Typically, one of the nightmares as a researcher is that you are presenting your results in front of an august audience, when somebody raises their hand and says “Have you looked at the paper by Jones, he does something very similar”, and you put on an embarrassed smile and say that you are not aware of Jones’ work! In research, you get no credit for reinventing the wheel – even if it is a beautiful, perfect wheel! So, you do need to spend considerable time and effort understanding the prior art in your field.

I must mention that some people directly jump into a problem and solve it, and later look around to see what is already done. I wouldn’t recommend it.

The literature survey step does two or three things: first, if you like the papers you read, you know that you are indeed interested in that area. Second, you get to know what is already done so that you don’t end up reinventing the wheel, and third, there are many modern techniques to solve problems that can only be found in research papers, so there is no way to teach yourself the techniques without reading recent research papers. Doing your research survey first is like putting the horse in front of the cart – the cart moves much more smoothly than if you tried doing it the other way around.

After the literature survey comes perhaps the most difficult step – to define a solvable problem. The keyword here, of course, is solvable. The problem that you define should be solvable not only in a global sense, but in particular, it should be solvable by you. This is where guidance from your research supervisor comes in handy, especially when you first start out on doing research (more on this later). The supervisor usually has a broad knowledge, and can tell what problems are likely to be interesting and solvable. If you work closely with your supervisor, then he or she probably also has a good idea about your abilities. So you would either define an interesting problem on your own, or define it with the help of your supervisor, and then start working on it. As you get more accustomed to the process of doing research, you will be able to define the problem completely on your own.

The third step is to solve the problem. Often, this will require you to look up references in diverse fields (especially in mathematics) to get hold of special tools to help you solve the problem. It is also customary to build a device or write a simulation software to show that your analysis is indeed valid. In other words, you have to be able to offer evidence as to why your idea works.

The fourth step is to document your results, i.e., to write a paper, to make a presentation, etc. Often, researchers underestimate the importance of this step, but selling your work is also extremely important. Just keep in mind that when you read a paper, you will find it interesting if you understand what the authors are talking about. Likewise, when you present your results to somebody, they will like it if they understand the point you are trying to make.

The four steps mentioned above take about an equal amount of time, at least initially. As you get good at writing papers, the fourth step can take less time, but it won’t be a quick write-up, ever. It takes time to write a good paper, to motivate your problem and why it is important to solve, to explain about past work in the area, to present your solution and then to summarize with evidence for the correctness of your solution.

I hope this description will help students make a judgment about whether or not they are cut out for research. Other points I can share are that you will like research if you like innovating or inventing, and looking beyond what is in the textbook or the problem at hand. Often, there is an impression that research is fun because you are always doing new things, while (for example) a software position is monotonous. That is not at all true. Research often involves a lot of grunt-work where you are hammering away at a problem or spending hours and hours trying to understand one little equation in some paper. Moreover, any position can be made interesting or monotonous depending on the attitude. I had an extremely well qualified and talented friend who once found himself in a verification/testing position in a start-up company. It was a tragic underutilization of his talent. However, he totally revamped the way testing was done at the company – he would never do the same test twice. He would do it once, and then figure out a way to automate it! In fact, he was doing research in how to efficiently do the verification/testing. So it is wrong to think that research is only done by people whose job title says “research”!

Also – do not expect yourself to produce world-class research results too soon. Research takes time! To give you an analogy, suppose you know nearly nothing about the flute besides the basic notes and how to play them. Suppose also that you wish to become so good at it, that you can play the flute in front of an audience consisting of Hariprasad Chaurasia and other great musicians like him, and have them say good things about they way you play. How long do you think it will take you to get that good?

Definitely, do not go into research because you have some idealistic notion that it is cool or fun or easy to do research. It is a lot of work, and you should enjoy doing the work. In addition, research requires you to be able/willing to think laterally, to be curious and want to come up with new ways of doing things, to work extremely hard, and to not give up. When Thomas Alva Edison said “Genius is one percent inspiration and 99 percent perspiration”, he was not kidding! In fact, not only genius, even average research is 99 percent perspiration!

In a more practical sense, if you found your undergraduate classes a drag, you didn’t enjoy the process of learning the new subjects in your curriculum, or if you found the going tough and the concepts difficult to understand, perhaps a research in the same field is not the best thing for you. If you really want to do research, I would recommend switching fields and picking up an area that you find interesting – it should pique your curiosity and make you want to learn more.

• Me: In a 4 year college term, when do you suggest is the ideal time for a student to decide upon his career – Research, Job, MBA etc., and if he resorts to research when do you think he should decide upon his topic of interest ?
• CM: I would say that during the third year is the best time. You would have had a flavor of courses in your department, and you can look ahead and find out a bit about what is in store for you. You can find out about the kinds of jobs your seniors have got, the kind of research positions available, and the prospects for getting into the MBA or the IAS. Yes, the IAS is still a good option!

• Me: Considering your field, a student of ECE has decided on research – how can he choose the right topic? ECE is an ocean with many applied applications too and every topic might be equally interesting and challenging for a freshman researcher – how do you suggest he take up the right choice which will suit him and which will help him ?
• CM: I actually don’t know the answer to this question. My best suggestion is to look back (and forward) at the different subjects and decide based on

(a) which subject interested you the most, and (b) which subject you did the best in. These two aspects answer your questions about which area would suit him (a) and which would help him (b). You can generally expect yourself to continue to do well in the area where you had an intuitive understanding and where you could do really well in your undergraduate study. For many people, (a) and (b) coincide, i.e., we find a particular topic interesting if it is easy for us to understand, and if it is easy for us to understand, we generally do well in that subject. But if the two are not the same, you will have to figure out how much weight you want to give to (a) and (b). That is, of course, dependent on the kind of person you are. If you are an “anything goes” type of person (which most of us are, actually), perhaps the specific area you specialize in won’t matter to you, so you might be better off choosing the area where you performed well in your undergraduate. There is no one answer that works for everybody.

• Me: Adjunct to the above question – Unless he works on a research/project in a particular topic in some field, he cannot know whether that topic is really of interest to him or not. Now how do u suggest that he goes about finding out a proper guide whom can guide him on a research topic in that field ?
• CM: Another good question. It reminds me of the story of Tenali Ramakrishna. Legend goes that when Goddess Mahakali appeared before him, She offered him a drink from one, and only one, of two cups: the cup of wisdom or the cup of wealth.

He said he wanted to taste both to see what they were like. Before She knew what was happening, he had gulped down the contents of both cups! I wish we could do that, to learn deeply about multiple topics and then decide which one to pursue research in. In an ideal world that is what one would do. However, we are constrained by time, and cannot spend years studying a subject.

There are many ways to answer the question. First, unless you are of the kind who simply cannot put your effort on something that doesn’t absolutely enthrall you, it doesn’t really really matter which field you specialize in. And if you do belong to the former category, hopefully you already know what field you ought to be working in!

The second is to talk to your friends and ask them for their candid opinion about where your strengths lie. You can classify your strengths into (a) analytical/mathematical inclination (b) ability to build things and © a knack for making things work.

Then you can talk to your undergraduate lecturers and see if they can guide you about what area to get into. Another good rule of thumb is whichever subjects fascinated you, or whichever subjects you found easy, are good candidates for your future research area.

As for the best way to pick a guide, there is nothing like a one-on-one interaction to make sure you can talk the same language and understand things in the same or similar way. If you can’t communicate with your future supervisor, it will be very difficult for you to do research under him or her.

• Me: Some may say that reading research articles will help him decide the topic of his interest, definitely a freshman cannot make much sense out of a research article, at that stage he won’t be able to completely extract all the information possible from a research article – so for him every article will be equally entising, again – what do you suggest that he can do to pin point his area of interest ?
• CM: There are some good magazines, such as the IEEE Spectrum or the IEEE Communications Magazine, or the IEEE Signal Processing Magazine

(there are similar magazines for every sub-field). These are partly meant exactly for this purpose – to help people who know little about the area get a better feel for it. So, reading these magazines would work better than reading some esoteric research article that may not make much sense you.

There is no end to reading articles if you don’t narrow down your area of interest first. As I mentioned above, unless you have a hidden, exclusive talent for something and no talent or interest for anything else, you can essentially pick any topic that seems interesting to you from the outside and take the plunge. If you don’t like it, you can always switch topics later. That’s one good thing about research – nobody will hold it against you if you switch topics. In fact, it is necessary, as topics mature and get “saturated”, you have to move to greener pastures. That said, just reading a bunch of papers in different topics, and not having any contributions at all, will not help you either. So once you pick up a topic, do work at it long and hard enough to at least make sure you are completely disenchanted before moving to a different one!

• Me: Is it always necessary to have a guide for research? Can a student go about on his own ?
• CM: Having a guide is useful, at least initially. The main things a guide can bring to you would be to point you in a fruitful direction of research.

However, keep in mind that what you do is not going to be research if your guide spoon-feeds you through the process of doing research. The goal for any student should be to eventually become an independent researcher, so if you are able to get there directly, then why not?

• Me: If a student is stuck at some point of his research, and he is not able to get the hurdle clarified by anyone, what do you suggest him to do at this junction?

Can he email profs in IISc, IITs or other schools asking his doubt? What are the chances of hearing back?

• CM: There are a few things you can do if you are stuck.

First, ask yourself why you are stuck – whether it is because you don’t have enough information or because the problem is basically difficult. If it is the former, you can try to look around and see if you can find a good reference that will explain a bit more about the problem you are looking into, and see if you can get some hints about how to solve the problem. Identifying when the problem is basically difficult or unsolvable is actually not easy – you will have to make a judgment call on it, and hope that you are right!

I would say that your chances of hearing back if you write to a faculty would be very good, provided that you phrase your question well and show that you have done your background work. Make a nice write-up explaining about the problem you are looking into, and show the steps you have taken to try and solve it, and finally explain why and where exactly you are stuck. Often, writing up this document will itself help you figure out the solution! Of course, the assumption here is that you will write to a faculty whose research interest matches with yours… if that is not the case, it is possible that you may not hear back. But, don’t take it personally, write to another faculty. Also, you don’t need to restrict yourself to faculty in the IITs and IISc.. you are welcome to write to anyone, anywhere in the world! What’s the worst that can happen? That the person will ignore your email, or write back asking you not to bother him or her, but that’s not a big deal, right?

• Me: Sure sir. There is certainly nothing to lose.
• Me: Is Research only about publishing papers and getting citations? Is there something more to research than just publishing scores of papers ?
• CM: Of course not. Publications are merely the way research is evaluated. I hope people doing research are doing it because they like the process of doing research, and the publications comes as a by-product, not the other way around!

• Me: What do you think is the most vital and unique part of Research that can never be got in any other spectrum of professions?

CM: As I mentioned above, do keep in mind that any profession can involve research, depending on how you approach it. So I guess the answer would be that the unique part of research is its uniqueness, that you are riding a wave that nobody has ridden before.

• Me: Some people work on topics that are completely theoretical, whimsical and have no practical application at all, what do u suggest on these research topics? Working for long hours, putting in loads of effort and proving something which has no practical implication, which can never be realized, which is just there only on the paper – do u think these research are sustainable ?
• CM: I’m guessing the person who wrote this question is not terribly fond of theoretical research! There are many research problems that have little or no practical application, but calling them whimsical would be to trivialize the researcher’s efforts!

Research does tend to be futuristic. If you could think of a practical application today, probably the research would have already been long done. In addition, not by everybody, and not all the time, does research with immediate practical application get conducted.

There is a place for every kind of research. There is a place for fundamental (or theoretical!) research, a place for practically oriented research, and even epsilon-delta research (where you extend existing results by a tiny amount). Therefore, the best way to look at it is what kind of research is most meaningful to you. What kind of research you enjoy doing the most.

Whether such research is sustainable or not, the answer is completely clear: yes. If you look back in history, any country that has done well, has done so because they have been ahead of their time. And they could get ahead of their time because they encouraged research. Advances in research are possible by a combination of advances in theoretical and practical problems. We should allow people who are extremely good at mathematics to come up with new results in mathematics without bothering about practical applications, and leave the finding of applications to other brilliant minds who are more suited to that kind of research. This, of course, makes it clear that research is sustainable only if it is done in sufficient “volume” so that all kinds of research can happen and feed into each other.

• Me: Why did you chose on coming into the academia over taking up an engineer post in any core company ?
• CM: Simple! I am happy being in the academia!

• Me: Sure Sir, Thank you sooo much for sparing your time and answering these questions. I really appreciate ur effort.

On behalf of Coffee With Sundar Readers, a very special thanks to you sir!