Tags Galore

Quotes From Here And There

Albert Einstein
Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction.
Dylan Thomas
I fell in love – that is the only expression I can think of – at once, and am still at the mercy of words, though sometimes now, knowing a little of their behavior very well, I think I can influence them slightly and have even learned to beat them now and then, which they appear to enjoy.
Eddie Cantor
It takes twenty years to become an overnight success.
Edward Abbey
May your trails be crooked, winding, lonesome, dangerous, leading to the most amazing view. May your mountains rise into and above the clouds.
e e cummings
To be nobody but yourself in a world which is doing its best night and day to make you like everybody else means to fight the hardest battle any human being can fight and never stop fighting.
Eyler Coates
We've all heard that a million monkeys banging on a million typewriters will eventually produce a masterpiece. Now, thanks to the Internet, we know this is not true.
Friedrich Nietzsche
Without music, life would be a mistake.
Gustave Flaubert
The one way of tolerating existence is to lose oneself in literature as in a perpetual orgy.
Going to the Opera is like making love; we get bored but we come back.
Johann Wolfgang von Goethe
If I love you, what business is it of yours?
John Steinbeck
Only through imitation do we develop toward originality.
Lady Mary Wortley Montagu
There is no remedy so easy as books, which if they do not give cheerfulness, at least restore quiet to the most troubled mind.
Leonard Cohen
Ring the bells that still can ring;
Forget your perfect offering.
There is a crack in everything;
That's how the light gets in.
The greatest thing in the world is to know how to belong to oneself.
Paul Sweeney
You know you've read a good book when you turn the last page and feel a little as if you have lost a friend.
Peter Altenberg
I never dreamed of being Shakespeare or Goethe, and I never expected to hold the great mirror of truth up before the world; I dreamed only of being a little pocket mirror, the sort that a woman can carry in her purse; one that reflects small blemishes, and some great beauties, when held close enough to the heart.
Robert Frost
In three words, I can sum up everything I've learned about life: it goes on.
There's no money in poetry, but then there's no poetry in money.
Satchel Paige
Work like you don't need the money. Love like you've never been hurt. Dance like nobody's watching.
Thomas Mann
A writer is somebody for whom writing is more difficult than it is for other people.
V S Naipaul
The writer has only to listen very carefully and with a clear heart to what people say to him, and ask the next question, and the next.

My Personal Library

On the JEE Brouhaha

I shared nanopolitan's parody piece Bombshell from IIT-K: Faculty Recruitment through JEE on Facebook, which led to the following response:

Q. How do you decrease the number of people below poverty line?
A. Simple, decrease the amount below which we call somebody poor.
Q. How do make everyone get into IIT?
A. Simple, ...
Its a clear political agenda where they found a completely wrong solution to a problem. An institute taking a stance in what it believes to be right is nothing to be parodied about and why would you? Are we really all that bored of the poverty line jokes on the government?

I just want to jot down some of my more serious thoughts on this.

1. I have a problem with the idea of a single exam and its numerical results playing so much part in the aspirations of a student. The notion, that the JEE is sacrosanct and that its results can actually tell if a person is intelligent enough or not, is ridiculous. I know many people will cry otherwise. (I have no biases for or against the JEE. I never gave the exam, nor am I greatly familiar with its format.) This is the reason why SAT and GRE are not factored in unilaterally in the admissions process of US universities.

2. The results of a single exam cannot be a better indicator than a result that factors in a student's efforts over a two-year pre-college program. If the method of incorporating grade 12 results is incorrect, biased or political, it should be corrected. But, the idea is correct in principle.

3. With the kind of competition, the winners are not very different from the people who almost made it. If you were to administer the JEE twice to the same bunch of students, there would be major shuffle at the top ranks. It would be better to bring in some more factors to have a holistic admissions procedure similar to US universities.

4. Years of preparation lead to freshmen being burnt-out when they enter the IITs. A question on Quora comparing students from MIT and IITs had answers which basically said that there isn't much difference intellectually; it's just that MIT students are more driven and passionate. See answers to Are IIT students smarter than MIT students? Success and happiness in the real world depends on personal qualities including but certainly not limited to intellect.

5. The fact that it is a decision based on numbers makes people believe that it has to be correct. This leads to disasters. See answers to the question How does it feel like to fail IIT JEE?.

6. It seems to me that the reason for such a furore over the exam is that people know that IITs are good only because they have an intensive admissions procedure and admit the very best students. This reduces the burden of education on the professors considerably. With the due disclaimer before generalizing and apologies to the few fabulous professors I have known at IIT Bombay, professors here do not seem to do as great a job as they are capable of. I suspect this is because it is an easier task to teach a class whose selectivity surpasses that of even the best universities in the developed world, and therefore the effort that makes great teachers is not put in. I am not sure if this is true at other IITs as well.

7. Simply put, out of the two aspects of rigor and possibility that education should impart, IITs focus only on the the former. This perspective spills over into the JEE as well.

IIT Bombay MTech (CSE) Alumni in Advanced Academia

A list of MTech (CSE) alumni from IIT Bombay who have pursued a PhD elsewhere. (There are many alumni of the programme who have liked the research atmosphere at IIT Bombay and continued with a PhD here.) This list is far from exhaustive, and has been written for incoming graduate students to have an idea of academic career prospects after IIT Bombay.

Carnegie Mellon University (CMU)

  • Avinava Kumar Dubey https://sites.google.com/site/kumaravinavadubey/
  • Bhavana Dalvi http://www.cs.cmu.edu/~bbd/
  • Kriti Puniyani http://www.cs.cmu.edu/~kpuniyan/
  • Meghana Kshirsagar http://people.cs.cmu.edu/Person/1374
  • Abhay Harpale http://www.cs.cmu.edu/~aharpale/

University of Illinois - Urbana Champaign (UIUC)

  • Manish Gupta http://www.cs.illinois.edu/homes/gupta58/
  • V G Vinod Vydiswaran http://sifaka.cs.uiuc.edu/~vgvinodv/

Dartmouth College

  • Umang Bhaskar http://www.cs.dartmouth.edu/~umang/

University of California - Irvine (UCI)

  • Vinayak Borkar http://isg.ics.uci.edu/people.html

University of Maryland - College Park (UMCP)

  • K Subramani http://www.csee.wvu.edu/~ksmani/

Rensselaer Polytechnic Institute (RPI)

  • Medha Atre http://www.cis.upenn.edu/~atrem/ (Post-doc at UPenn)

Pennsylvania State University (PSU)

  • Neela Sawant http://www.personal.psu.edu/nks125/


  • Shirish Phatak http://www.research.rutgers.edu/~phatak/

Indian Institute of Science (IISc)

  • Rupesh Nasre http://users.ices.utexas.edu/~nasre/ (Post-doc at UT Austin)

The Machine Learning : Pattern Recognition : Information Retrieval : Artificial Intelligence : Natural Language Processing research group at IIT Bombay consists of the following faculty members:
Find research interests of all faculty members here.

The Best Nurturers in Computer Science Research

Recently I came across a paper quantifying success in the intangible process of mentorship in computer science research. The work was done around 2004 at IISc by Bharath Kumar M. and Y. N. Srikant and published in Proceedings of the Fifth SIAM International Conference on Data Mining. [Technical Report] [ICDM Paper]

Not surprisingly, the paper states that "there is a recognizable deviation between the rankings of the most successful researchers and the best nurturers, which although is obvious from a social perspective has not been statistically demonstrated."

Interestingly, though the work mentions quite a few Indian nurturers, almost none of them have consistently resided and researched in India.

helo sir

Every couple of months, I must suffer the (somewhat amusing) trauma of an email like this one in my inbox:

hello sir,
           hope u r fine....i m also....my name is Aaa bbb
and  I m doing my m.tech from computer science  from aaaa kkkkk
engineering college, its my 2nd sem.....i m not able to decide my
project,actually i have no interest in coding and so i dont want such
project in which coding is required...so please can u suggest me any
project.....need ur help please

Though I generally try hard not to judge people, there are any number of reasons why an email like the one above is wrong. A masters student in CS with no interest in coding?! Further, since the person is doing a masters and has no interest in coding, (s)he may end up teaching in one of our many "educational" institutions, which depresses me no end.

I cannot even begin to imagine the pain (multiplied many times over) that faculty at IITs must suffer handling such emails almost everyday. See this to know what I am talking about.

Sudden Debt: A Blast From The Past

Sudden Debt: A Blast From The Past: "I have been re-reading Gustave LeBon's The Crowd , the 1895 classic on the psychology of herd ('crowd') behavior. The French social psycho..."

On Music, GRE-TOEFL, And Other Sundry Things

I find that I am increasingly obsessed with Indian (Semi-)Classical and spend hours listening to beautiful, brilliant songs. Perhaps, my only true wish currently is to learn some of this amazing musical legacy. And in a hypothetical exchange of fortunes, I would readily trade everything I have for mastery of this most beautiful of artistic forms. Ustad Bade Ghulam Ali Khan once said, "If, in every home, one child was taught Hindustani classical music, this country would never have been partitioned." Of all generalizations in the world, I believe this to be most likely true.

On June 29, I gave the GRE - an exceptionally inane test that claims to gauge preparation for graduate studies. Quantitative was a cakewalk as expected. I lost a few points in verbal to some boring RCs. Finally ended up with Q:800 and V:680 Enjoyed writing essays after a really long time. I have planned to give this exam for some time now. Glad to have finally finished it off. Don't tell me new year's resolutions don't work!

I had a research section in the end in which I wrote an argument essay. Perhaps, the ETS wants to move the essays to the end of the test eventually, and wants to obtain statistical estimates of whether the scores in essays change if the section is moved to the end of the exam. Though the GRE is a fairly mundane test, the assiduity in maintaining the score norms of the standardized test is commendable, unlike in India where it is quite difficult to find policy decisions backed by hard statistics, especially in education.

I have recently shifted residence to the highest floor of Hostel 14, IIT Bombay, along with eight of my friends. My new accommodation has a nice view of the Powai Lake and the Hiranandani skyline. It's a nice place to spend the next one year.


Update (03/08/2011): Gave the TOEFL on 23 July. Scores came in today, almost a week before they were expected. With a decent score of 114, I can now tick off another of my new year's resolutions! Finally at the beginning of the end of apping...

Advice For People Considering CSE, IIT Bombay

Update (28 September 2011):

From recent personal experience, some of my statements about placements in this blog are not true. Some of the consulting/finance/other-rubbish profiles are off-limits for graduate students even before any elimination rounds are conducted. Hence, if you are an ambitious careerist interested in being in the industry after your graduate studies and have admits from IIMs or good US/European universities, you might want to give IIT Bombay a miss for greener pastures. After almost a year and a half at IIT Bombay, I have realized that the institution is good only for academic rigor and some of the amazing professors that teach here (my favorites here). Such professors are exceptions to the norm but the situation might be better than anywhere else in the country.

The administration is thoroughly Indian; it sucks as usual. One of the amazing professors I mentioned in the earlier paragraph (Prof. Soumen Chakrabarti whose work has been cited in the PageRank/Google paper http://infolab.stanford.edu/~backrub/google.html) put up the following description of his 'Web Search and Mining' course on IIT Bombay's Course Management System: 'Main building still knows this course by the mouthful "Information Retrieval and Mining for Hypertext and the Web" in spite of requesting them to change the title over five times.'

Recently, there was a major case of food poisoning in three of the hostels affecting more than 1000 people in all, and nobody has taken up even moral responsibility for the fiasco! The concerned caterer has been fired and a new caterer will take his place this semester.


A query I received recently about studying at CSE, IIT Bombay:-

Dear Abhinav,

My name is Axxxx and I just got an AIR of 44 in GATE 2011 which means that I will probably get into the MTech CSE at IIT Bombay. I got your contact off the IIT website and I was wondering if you could help me out by sharing some of your experience. Your profile seemed interesting and I know I can trust somebody who gets XKCD.

So, how do you find the course compared to your BE at VJTI? (I'm from TSEC and really need a change in the level of teaching standards). What about life at IIT campus? And do you know anything about the placements scenario for MTech students? Is it on par with the BTech placements or is it not that good? I have an offer from Oxford University but I'm probably declining it as the UK visa rules are getting extremely annoying for anybody who wishes to gain work experience after their degree there (plus its bloody expensive!).

Anything that you can tell me will really help me in making this decision (IIT vs Oxford) better for me...

Regards,A B

My response to the query:-

I think I will set out the reasons for my decision to come to IIT Bombay. You must evaluate your situation, opportunities, and inclinations to come to your own decision.

I am fairly certain that I will end up either in academia, an R&D lab, or doing a startup. I joined IIT Bombay because I needed a stronger background to pursue a PhD, both in terms of my technical soundness and the reputation of the institution I was going to study at. I had admits from Uppsala for technical masters, and IIM Lucknow and Kozhikode for MBA, besides the IIT admits. I chose IIT Bombay because I was sure I would eventually get bored with standard management stuff (is that really work?), and (surprise, surprise!) I still liked CSE after I graduated from VJTI. I haven't really regretted my decision so far. IIT Bombay provides fabulous facilities and a beautiful campus, and is about an hour from my home.

VJTI, the institution where I spent my undergrad days didn't have a research culture to speak of. I don't really remember any worthwhile professors or well-taught subjects. In short, my undergrad education would amount to zilch for a person looking for rigorous foundations. Consequently, the difference between the UG/PG quality of education has been immense for me. At IIT Bombay, I have been taught and guided by some really great professors; some of them are forerunners in their particular areas of research. Prof. Bhaskar Raman (Systems and Networks) and Prof. Sunita Sarawagi (Machine Learning and Data Mining) are my personal favorites. Prof. Soumen Chakrabarti is also highly regarded and really knowledgeable, but he did not click with me.

The graduate intake at CSE, IIT Bombay is much larger compared to other IITs. This is a good thing, because it leads to bigger research groups and you almost always have people who share similar research interests. It's a bad thing, because the class sizes increase beyond the desirable graduate class size (maximum ~20-30), leading to lesser attention of professors to each individual student. For better or for worse, the competition for getting into research groups and working under good guides also increases.

Also, IIT Bombay has a much more vibrant social life than other places like IISc and other IITs. It has great placements and a good reputation in case one wants to pursue research further. As far as I know, IIT Bombay has had good placements, but individual preparation and commitment are better indicators of the quality of placements one lands up with rather than past statistics. From hearsay, the placement opportunities for MTech guys are at par with those for BTech guys except for some selective I-banks.

Correspondent's subsequent reply:-
Dear Abhinav,

Thanks for all that information. The one big factor for me wanting to choose IIT over Oxford was that there is very less research activity on distributed systems and machine learning at Oxford (they focus more on theoretical issues of computer science) but I found quite a few active projects at IIT. The only doubt in my mind was whether the MTech degree was good enough or not.

I have seriously been encouraged to go for IIT after hearing your story (and that of a couple of others). For some reason, I had an ill-formed view that MTech at IIT was inferior to the BTech degree (probably wrong feedback from some IIT undergrads).

I had read Prof. Raman's web pages before, and some faculty in my college were telling me about one Prof. Kavi Arya as well. Knowing that professors like these will be teaching us is really exciting.

I'll hope to meet you in a few months then (assuming of course that I get direct admission)... Good luck with your exams and stuff!

Thanks again.


My subsequent response:-
Hi Axxxx,
Due to lower competition, the quality of MTech batches in most departments at IITs is lower than BTech batches. That does lead to the currently prevalent perceptions.
However, for CSE, the keen competition does lead to a pretty decent batch, at least at IIT Bombay. I know of seniors who have been placed in Google, Microsoft, Oracle, Intel, Credit Suisse, etc. or have gone to decent universities like CMU, UMD, etc. for their PhD. It really depends on the individual caliber of the student.
There is a difference between the BTech and MTech batches, because the BTechs receive rigorous training in fundamentals for four years. More importantly, BTechs have healthier attitudes toward learning and often ask a lot of questions in a class making it interactive and worthwhile. These are differences that cannot be wished away, but which many MTechs overcome through their first semester.
Regarding admissions, you most certainly will get a direct admit to IIT Bombay with AIR 44. Apply to two other IITs of your choice as a backup. I applied to all of them and never had to bother with admits from the other IITs.

Abhinav Maurya

Advice for new matriculants:-

Hi All,

I hope you have settled well @ IIT Bombay. To those of you who contacted me earlier, my apologies for not having met you until now due to some other commitments.

The first semester at our department is a very hectic and important one. In most of the cases, performance in subjects of the first semester determines your seminar in the second semester, which usually leads to an MTP in the second year. The topic of your seminar and MTP need not be the same. So, the plan usually is exploration of CSE research areas and groups in the first semester, and consolidation/taking advanced courses in the second semester, finishing with the grand MTP later. In conclusion, your choice of courses now decides much of what you will be doing at IIT Bombay. So, please consider your choice carefully by attending as many lectures as possible until the registration closes.

Though there is the possibility of taking up seminar in the first semester, it is conventionally taken up by students in the second semester. This gives you a semester to familiarize yourself with the flavor of CSE research carried out at IIT Bombay.

Each research group often has introductory graduate courses whose grades are often used by professors in choosing a student for seminar/MTP. These should be clear in the fundae session to be organized soon.

There are certain "load-balancing" courses in the department. If you are really interested in CS and your reasons for coming here extend beyond the placements and the IIT tag, I suggest you to take courses that interest you and which are moderately difficult as per your current preparation. Your MTP is worth 90 credits and far outweighs the 60 credits of the courses you will take. Besides, liking a course makes it far easier to do better in the course.

For people interested in studies beyond the conventional CSE stuff offered in the department, you may want to explore Mathematics, Applied Statistics and Informatics, EE, and IEOR courses in later semesters. These departments often offer courses in niche areas that are closely related to CSE.

On deciding courses, you can take two courses offered by two research groups that interest you. For example, Foundations of Machine Learning and Artificial Intelligence for AI-ML-Data-Mining people, Advanced Computer Networks and Network Security for Networks people, Program Analysis and Functional Programming for Systems and Compilers, etc. If you are completely undecided, you may wish to explore four different areas by taking one introductory graduate level course in each area. The final mix of courses is for you to decide. Kindly discuss with facad/any professor/me/any senior if you feel you need to talk about your choice of courses.

I hope to talk to you all at least once before the registration closes. On Monday, I will be in Circular Hall from 10:30 to 12:30 and then again from 15:30 to 17:30. If you are at KReSIT, please drop by. Please feel free to bug me at any point during the semester for any reason at all.

Have a great time at IIT Bombay. I am sure you will! :-)

What is it with the unknown?

Why does one find oneself greatly fascinated by something that one does not even understand?

My joy and fascination with Indian/Western classical music remains incomprehensible to me, given my scant knowledge of the subject. Yet, I find myself listening to (semi-)classical music all the time.

My fascination with literature has waned the more I have known about the "tricks of the craft". Now I find it tedious to read simpler "story" books and expect something as good as Kafka every time I pick up a book.

IIT Bombay: CS 699: Software Laboratory

This is a blog about the Fall 2010 offering of CS 699, the Software Laboratory course @ IIT Bombay.

Update 27/08/2010: This course is a course which makes little sense to me. It aims to teach us LaTeX/Beamer, Python, Shell Scripting, and Web Design (HTML/CSS/Javascript). It is not understandable how an institution comparable in quality and reputation to some of the best institutions around the globe can place such a ridiculous course at the graduate level, and worse still, assign it credits more than any other course in the department (our normal courses carry 6 credits). The course carries 8 credits which is a high number by our credit structure. I do not understand why the course cannot have lesser credits or be split across the two semesters allowing more time for the assimilation of the skills that it attempts to inculcate in us. The course can also be structured as a month long audit primer that graduate students take prior to beginning their first year CSE electives. In a more general sense, it is a manifestation of the rigidity which is often found in Indian educational institutions. The blanket assumption that these skills are necessary and must be shoved down our throats in a single highly loaded semester is beyond my comprehension. Here's a past student's view on the subject.
This is a blog about the Fall 2010 offering of CS 725, the Foundations of Machine Learning course @ IIT Bombay.

Update 27/08/2010: The course is taught by Prof. Ganesh Ramakrishnan who is offering this course for the first time. Last year, it was offered by Prof. Sunita Sarawagi. Though the professor knows a lot about the subject, the lectures could do with some more organization. In particular, the planning of the course over the semester is still sketchy and motivations are vague most of the time. Do not let the word 'foundations' deceive you; it's a particularly heavy course and needs you to know probability, linear algebra, calculus, and statistics at a fairly high level to appreciate the points that Ganesh makes in the class. This is a prerequisite for some advanced courses in Machine Learning offered by the department. The CSE department @ IIT Bombay is known for its Machine Learning research group, and this is the course that teaches you to make sense of the basics of the field.

Copyright © 2007 | All Rights Reserved | Best Viewed In Google Chrome With A Resolution Of 1024 x 768