Chances in Information Extraction(Machine Learning) in top 20 schools.

[TextMiner]

I am currently a Research Assistant at IIT Bombay working in the field of Information Extraction / Machine Learning. I finished my Masters in Technology from IIT Bombay in 2010 and have a couple of publications(not as a first author though) in a Tier-2 International conference in a slightly different field. Having applied in the top 9 schools during previous Fall app. process, i was severely demotivated by the rejects received from EVERY ONE of them ; (It is possible that my publications were not in a reputed conference ; Also a very famous recommender whom i approached might not have had a favorable opinion of me reflecting in my rejection ; Again a speculation).

Since the fall application for 2011, i have done an Internship in a reputed Search Org. The work that will result out of this is currently being shaped for submitting to PAKDD or SIGMOD. Also currently i am undertaking a Research assistantship under a leading authority in Information Extraction in IIT Bombay.

For the Fall 2012 season, i am expanding my applications (an intelligent selection of max. 10) in the top 25 colleges. My main interest is in the general area of Machine Learning with a focus on Extraction problems especially large-scale extraction. Can someone please let me know my chances for a PhD in the top 25 as mentioned ? Any addendum to the information i requested is perfectly fine, for example, why UC Irvine is a good college to apply though it might be out of the top 30 etc.

Thanx in Advance.

[starmaker]

Your chances seem fine to me if you have good recommendations/statement of purpose/fit with the department, but if you really want a PhD (i.e. you'd rather get one from a lower-ranked school than not get one at all), you should apply to some schools not in the top 25 as well. The admissions process at the top schools is very competitive, and you simply can't assume that you will get into one even if you are a very strong applicant.

I don't know the specifics of which programs are good at information extraction, but I can give you some other reasons besides chance of acceptance for expanding beyond the top 25 too:

- A program that is lower-ranked in computer science in general might be higher-ranked in AI. Some departments have subfields that are much stronger than the rest of the department. People who hire PhDs in AI fields will be aware of which programs are good in AI specifically.

- A program that is lower-ranked might have a particular faculty member in your field who is very strong and well-known, and this faculty member may be able to place you into a job, after you get your PhD, that is better than what graduates of the program normally get. One of my prospective programs is like this - it's not known as a strong program in general, but it has one very famous and well-regarded faculty member who happens to work right in the area that I want to study.

- If you are one of a program's stronger admits (because it is lower-ranked and there is not as much competition for admission), they may offer you significantly more money to study there.

[TextMiner]

Hey, first off, thanx a lot for responding. Appreciate that. Let me take each of your points.

Your chances seem fine to me if you have good recommendations/statement of purpose/fit with the department, but if you really want a PhD (i.e. you'd rather get one from a lower-ranked school than not get one at all), you should apply to some schools not in the top 25 as well. The admissions process at the top schools is very competitive, and you simply can't assume that you will get into one even if you are a very strong applicant.

It seems a valid point ; Of course, my SOP may not explicitly highlight Information Extraction, but i will bring it out in the course of explaining my research. I will broadly put my interests in Machine Learning. Do you think its a fair point ?

I don't know the specifics of which programs are good at information extraction,

There are a number of good papers churned out from not only the top 20, but even among the top 35 (am going by the US News & Reports rankings, which may be biased to an extent). We could take it offline where we can discuss on this.

but I can give you some other reasons besides chance of acceptance for expanding beyond the top 25 too:

- A program that is lower-ranked in computer science in general might be higher-ranked in AI. Some departments have subfields that are much stronger than the rest of the department. People who hire PhDs in AI fields will be aware of which programs are good in AI specifically.

- A program that is lower-ranked might have a particular faculty member in your field who is very strong and well-known, and this faculty member may be able to place you into a job, after you get your PhD, that is better than what graduates of the program normally get. One of my prospective programs is like this - it's not known as a strong program in general, but it has one very famous and well-regarded faculty member who happens to work right in the area that I want to study.

- If you are one of a program's stronger admits (because it is lower-ranked and there is not as much competition for admission), they may offer you significantly more money to study there.

Some very interesting points raised ; The professor whom i am doing my assistantship under, informed me about a downbeat situation in the US regarding funding when she was there for KDD. So yes, funding is definitely a concerning factor for me. I understand that the funding is Professor-specific more than being Department specific. Do you think in that case it would be wise to contact the professor of the so called lower ranked university in advance ?

[starmaker]

RA funding is professor-specific (except in programs that have rotations, which CS programs don't usually have - in those, the department usually funds student rotations, I think). TA funding is departmental in many departments. Internal fellowship funding can be through the department/program, school, or university, depending on the fellowship.

"Do you think in that case it would be wise to contact the professor of the so called lower ranked university in advance ?"

Yep. If anything, professors at lower-ranked programs are more amenable to chatting with prospective students about their research and their program than many profs at higher-ranked ones (who are deluged with attempts by prospectives to make contact and gain favor, and may get tired of it).