First Semester (Fall 2004):
Algorithms and Complexity
Instructor: Prof. Sundar Vishwanathan
A wonderful course. Sundar's sense of humour, style of teaching, the beauty and creativity of those NP-completeness proofs, and thoughtful assignments made this course one of the best I have had here. The only sad part was the last 30% of the course, where he started Linear Programming, leading up to the Simplex method. Historically, I have been bad at Linear Algebra, and this was no different, and I suffered, through the class, and through the exams - ending up with a very bad grade. But still, I loved this course for the first 70%.
Formal Specification and Verification of Programs
Instructor: Prof. Supratik Chakraborty
Term paper
This was a surprise package. I had no idea what Formal Methods were all about, and by the end of this course, I could nod my head sagely when people mentioned formal verifications, model checking, 2 power 20 states and beyond, protocol checking, etc. Supratik is an incredible teacher. I wish I had had the money/guts to take a handy-cam to his class and record his lectures. They need to be preserved. Super-earnest, dedicated, lucid, thorough, student-friendly - he is all I could ask for in a graduate level teacher. The only sad part about the course is the lack of reading material on the net/elsewhere. This is the only course in my entire academic life where I have attended all classes. If you are from some standard run-of-the-mill university in India, you might not have been exposed to the world of Formal Methods, which is a very active area on theoretical computer science research - do take this course.
Probability and Statistics
Instructor: Prof. Shashikant Kelkar
A course with a lot of potential, but thoroughly wasted by the instructor. Its quite sad that this is a mandatory course for SIT students. I will blame my lack of appreciation for probability and statistics on this course, and more so, the instructor.
Foundations Lab
Instructor: Prof. Kavi Arya
Absolute waste of time, and no marks for guessing that this is a mandatory course as well. The teaching assistants did not have a fan-following due to the inevitable nature of their duties, and to add to the irony, in my second year, I am one. This course attempts the impossible: to teach people how to program in all environments, in all flavours, to get them to learn all the tricks of the trade, and other superlatives. It partially succeeds to the effect that people who have never programmed anything above quick sort might end up being more confident about writing web applications in php. So what? I ask. I believe that people will learn programming when they need to and with this Foundations Lab based system, we aren't giving ourselves enough credit. 10 credits wasted on this one.
[Oct. 2010 Edit] - In retrospect, I have to admit that this course did serve its purpose. Which was to instill confidence in students that they could build non-trivial systems under tight timelines. Invaluable skill, that. I eat humble pie now. [Oct. 2010 Edit]
Second Semester (Spring 2005):
Approximation Algorithms
Instructor: Prof. Ajit A Diwan
Diwan is the archetypical super-genius theory professor. Very fast on the board with the chalk. Very thorough. Almost intimidating in spite of his dimunitive physical presence. The course is brilliant. Again, the lack of material sometimes hurts. But with Vazirani coming up with a student friendly text book, this course is do-able. Diwan covers a whole range of problems, and does a very matter-of-fact job in the class. Scheduling theory stuff is very cool, and so is the primal dual stuff. Just that I couldn't understand it enough to appreciate its beauty. So, a bad grade was expected. Blame it on lack of work, and lack of assignments. Should've spend a lot more time on the assignments. But all in all, a fantastic course, very practical, and of course, its worth the experience just to see Diwan in action.
Hypertext Retrieval and Mining
(Course page)
Instructor: Prof. Soumen Chakrabarti
The only problem with this course is Soumen's voice/diction and the seeming hodge-podge nature of the course content. Otherwise, this course is very interesting. Shows the web to be what it truly is. His textbook is also quite a good read. The course covers some classification, clustering, semi-supervised stuff, PageRank, Hits, advances in ranking methods, SVM optimization, and a whole load of other related material. There was some disturbing math in the middle of the course, where he seemed to be exploring some new body of work along with us in class. But with this area evolving at such breakneck speeds, I guess its inevitable that he teaches stuff which he doesn't fully understand (at that time). Also, he tried to cram in a lot of basic tools together with applications and advanced material. I guess that is being corrected now with the Statistical Foundations for Machine Learning course he is offering. With basics covered there, maybe Hypertext and Text mining will take the entire course in the spring semester. One thing to watch out for in this course is the ideas that Soumen keeps throwing at the class as if they are offhand comments. Each such idea is worth a project at worst, or a paper at best.
Data Mining and Data Warehousing
Instructor: Prof. Sunita Sarawagi
A good introductory course to Data Mining, and very little of Data Warehousing. Covers the basics: decision trees, regression, neural networks, probabilistic learning, active learning, association rules, deduplication, etc. Sunita is very matter-of-fact in class. But at some level I sensed that she was never able to connect with her class - for one, I never was able to truly appreciate her desire to make us see real world applications. She was quite fastidious about data sets, open-book vs closed book exams, expected answers, and the like. Being such a basic level course, I expected a lot of aesthetically beautiful God's-Book kind of material in this, and didn't find any such thing. On the good side, she covers a lot of material, some of which is her own research, which has a very commanding feel. This is probably the best data mining course in India.
Applied Economics (Graduate Institute Elective)
Profs. Ramanathan and Narayanan (Dept. Of HSS)
A graduate level elective is mandatory for all Masters students in IIT. This elective is supposed to expose us into other domains and expand our horizons. I took this particular course with great enthusiasm, expecting to learn about demand, supply, prisoners' dilemma, tragedy of the commons, law of comparative advantage, and other universally applying economic principles. Instead, I just learnt how not to antagonize old-school professors, how to fare better at reading comprehension, and a very little bit of economics. Most of the course was basic knowledge of the workings of the world. Not that I am saying that its any easier. But, this course didn't give me anything more than I already knew from basic non-economics related undergraduate level study. The formal part of economic study was almost non-existent. The latter half of the course was redeemed by the second instructor who was reasonably good. Again, this course had a lot of potential that just went wasted.
Systems Lab
Instructor: Prof. Bernard Menezes
This course attempts to give students a feel of real world software development, with a full semester length project which tumbles through an ugly waterfall. This course's model is as volatile as the software development model, changing with each offering, trying to learn from past mistakes, moving from monolithic teams to small atomic teams, and other experiments. For people who have never worked in the conventional software industry, this might be useful. I am not sure of it, but that seems to be the good intention behind this course. Teams of 2-4 pick a project, and take it through the waterfall. Again, absolutely no credit for guessing that its a mandatory course. Ah, now you sense a pattern? ;-)
Third Semester (Fall 2005):
Statistical Foundations of Machine Learning
Course page
Instructor: Prof. Soumen Chakrabarti
Fourth Semester (Spring 2006):
Approximation Algorithms (sit through)
Instructor: Prof. Sundar Vishwanathan