IIT Madras M.S. (by Research) Admission Test and Interview in EE1 (Communication)

[This is an old post from 2011 and some information may be outdated, but you'd get an essence of how these interviews are conducted.]

If you think you are a bit passionate about your subject by the end of your engineering, perhaps the easiest way to get into an IIT is to get an MS admission. It doesn’t require a very high GATE score. You are eligible to write the admission test if you just meet the norms (for us, it was 65% in UG and 550 GATE score; and a very good score in one can offset a poor score in the other).

It is to be noted that MS needs a research-oriented mind. This doesn’t mean that MS is only for those who want to do a Ph.D. after its completion. An MS student will have as good opportunities for joining industry as for continuing his/her research.

Here, I will be concentrating on the EE1 (Communication and Signal Processing) test and interview at IIT Madras.

Syllabus

The test will be in the areas of Signals and Systems (all of Transform Theory, Fourier Series and Sampling), some concepts of Signal Processing (mainly DFT) and Probability Theory and Random Variables. The test now includes Linear Algebra as well (see Updates). (The website also mentions Electromagnetic Field Theory, but I have never heard of questions coming from that subject in the exam.)

The question paper

The question paper will be of problems, some on signals and relatively fewer on probability concepts. Answers should be descriptive. Each question will be simple, and generally would not require much time to finish, if the idea of what to apply just strikes you. Note that the evaluation is based on the method we have used to solve the problem. Minor calculation mistakes are usually overlooked, if the procedure we adopted is correct. So, in our solutions, we need to be clear in mentioning the steps involved to get the answer.

Preparation

1. The best book for Signals and Systems is Oppenheim (Signals and Systems, Oppenheim & Willsky). Solving the exercise problems would help. In the exam, much importance is given to the following:
1. Properties of transforms: The question will be very easy; we need to use one or two properties of transforms to get the answer. If we get to know which property to apply, the problem is done.
2. Relation between Fourier Series and Fourier Transform: Most often, the question contains simple manipulations on those signals which have both Fourier Series representation and Fourier Transform. We need to understand how they are related to each other.
3. Relation between Fourier and Laplace/Z transform for continuous/discrete time signals.
4. ROC of Z-transform: Causality, stability of system based on ROC, and significance of poles and zeros in it. Questions can be like - what happens if we modify a pole, or if we rotate the axes by 180°, etc.
5. Sampling: Aliasing in sampling based on Nyquist rate, choosing suitable sampling frequency for a particular input, etc.
2. DFT can be read from any of the standard books, e.g., DSP by Proakis and Manolakis.
3. Probability and Random Variables can be read from Papoulis. Alternatively, some faculty members of the department have uploaded assignments (and their solutions) as part of their teaching coursework in their web pages. These could really help. Importance in the exam would possibly be on the following topics:
1. How to apply Bayes theorem.
2. Differences between PDFs (continuous) and PMFs (discrete), and how to determine probabilities from them.
3. Determining conditions in which a given PDF is valid, or checking whether a given PDF is valid.

The interview

We need to pass the test to go to the interview, which will be on the same day as the test. The most important criterion in the interview is that some clues will be given regarding the mistakes we made in the test, and we are expected to catch the idea and correct the mistakes. So, it is better, before going to the interview, to discuss all the answers and know the correct solutions. We need to be confident in what we are saying.

If we have mentioned some specific areas of interest in our research proposal, we will be asked questions on them by the relevant faculty. However, we are not expected to know too much about them; a minimum knowledge is sufficient. If we were quite general in our proposal, then we will be asked about our preferred specialization and related questions, and we may be given a research area according to their requirement. Some questions on our UG final project will be asked, if we are interested or if it is relevant.

Updates

The test now includes linear algebra. YouTube has very nice videos of Gilbert Strang's lectures. Or read his book Linear Algebra and its Applications.

I added a post on how to write a research proposal. However, for MS in IIT Madras, the proposal need not be as rigorous as I described. If you clear the written test and if you haven't mentioned a specific topic in your research proposal or in your interview, then the faculty will decide your adviser, depending on your performance.

On-campus placements: Most companies registering for them have generic requirements of the candidate's profile. While some scholars register and get placed on campus, some look for job profiles specific to their research areas. They apply individually to the companies they desire, usually through their seniors or contacts from their advisers, and get placed. Some convert their master's to a Ph.D., or apply abroad.