User:Iitk06

This is a general page where the seniors post information as to where to app and information about universities.

Berkeley http://www.eecs.berkeley.edu is probably one of the best places to be given that you get the best climate here, enjoy the bay sitting right next to one of the most beautiful cities in us, san francisco, the amazing variety of food here and the cultural vibrancy. Its notoriously known for its high level of student activism (free speech movements and others).Berkeley is the only university in the nation to achieve top 5 rankings for all its PhD programs in those disciplines covered by the US News and World Report graduate school survey [:-)] which makes interdepatmental research very active here.

Changing your area is easy in berkeley as you are given a year to decide your advisor, which is the best part. Not too much force on coursework and you get a lot of time to explore. Do not constrain yourself to apping in a particular area. For eg. theory group is excellent here but they generally do not take many students. so its a good idea to app in for eg. theory + networks at berkeley to increase your chances to get in. Also if some of you are interested in computational biology then consider looking at http://computationalbiology.berkeley.edu. Its possible to so a CS phd with this dept as some of the CS pofs have a joint appointment with this dept.

Theory: One of the best places to be. Richard Karp, Christos Papadimitriou, Luka Threvisan are well known but as I said they dont usually take students.

Systems, Networking, Databases The best systems group in US. Lots of excellent profs and projects which are tied to the industries like google, intel, microsoft. Being near silicon valley lots of opportunities for joining startups too. Beware the average time you spend on a systems phd is a year or two more than a theory one!!!

Ai, Vision, NLP: One of the best vision groups here. Most of the benchmarks in Object Recognition are held by berkeley;-). This is a major area of thrust here. Though Forsyth has moved to UIUC a lot of collaborative works goes on. The Optometry and neuroscience depts are also good here and vision research here has a slight bent towards biologically motivated systems.

Statistical Learning Theory: Mike Jordan, Peter Bartlett and others. Lots of students!!

Natural Language Processing - Dan Klien - new professor with lot of enthu and others

Not so good things here Robotics : So So ..go to CMU! (muwwaaahaahaaa! see below, ye innocent lambs!), Graphics: UNC.

For others see http://www.eecs.berkeley.edu/Research/Areas/

Carnegie Mellon http://www.ri.cmu.edu

Great place to be for Robotics, Comp Sc, EE (a new thrust on power systems from what I hear), Neuroscience, Fine Arts and Drama (NLP and Speech(?).

What I like best: Qualifiers do not include writing exams in random fields you ought to have 'breadth of knowledge' in. The 'IIT' name is very well-known, amongst students and faculty- that makes me both happy and nervous.

In particular, robotics:

Vision: If the name Takeo attracts you, do not come. He's a nice guy, enjoys a joke, does fantastic work, is a demi-god (or maybe, you can scratch out the 'demi'), is getting on in life but is unstoppable as ever, and is decreasing his student intake each year.

Field Robotics: One of the best places to be is here. The field robotics center is huge and there's a great deal of work going on all sorts of stuff. Be warned however that jobs after that can get hard to find- for jobs in NASA, JPL etc, you need to be a US citizen, for academia, there are very few places that can afford to run high quality field robotics research labs.

Language Technologies Institute- (distinct from RI)- good place for NLP- very selective as far as PhD goes, but take in Masters students and all the Masters students who came in this year had to pay for the first semester, but managed funding for the rest of the year.

Human-Computer Interaction- another up and coming area- combines Design, Psychology, Computer-sciency stuff. This is where you talk about what is a good way to get electronic teaching aids to schools without antagozing teachers or putting students off, or what gestures made by a Honda ASIMO will make a human react to it as he does to another human vs a 'silly machine'.

To be contd.

Johns Hopkins University http://www.jhu.edu

Okay if you want to come to JHU you want come here either for Biomedical Engineering (pretty much the best in the US and the world) or for Speech and Language Processing (JHU and CMU are loggerheads about this in more than one way!). Other areas are only averagely well known here.

Biomedical Engineering: Pretty much the biggest thing in engineering here. Also research in other areas is often in many ways related to Biomedical due to the funding. (Rest to be filled by more knowledgeable persons ...)

Language and Speech Procesing: JHU houses the Center for Language and Speech Processing http://clsp.jhu.edu which is one of the bigger federally funded labs concentrating on this area. The funding situation is pretty good for now (they took in six new people this year) but do contact some senior there if you wish to apply next year (i.e 2007). The center takes students from both CS and ECE and people are generally interviewed by one or all faculty of the center before they are accepted.

In particular here are the respective faculty and their general reserach interests:

Frederick Jelinek: Information Theory, Statistical Speech Processing, Language Modeling. He is pretty much the founder of modern Speech Processing with fundamental contribution in Information Theory (BCJR coding) and Statistical Speech Processing (HMMs and their ilk). Though he is the director of CLSP, he has reduced his student intake.

David Yarowsky: Word sense disambiguation, Minimally supervised induction algorithms in NLP, Multilingual Natural Language Processing, Information Retrieval

Jason Eisner: See this http://www.cs.jhu.edu/~jason/research for the best description

Sanjeev Khudanpur: Information theoretic methods to human language technologies such as automatic speech recognition, machine translation and natural language processing, Content based Information Retrieval.

Note that the alumnus of CLSP are well-placed in the various companies and research labs in US and hence have a strong network across the institutions.

Massachusetts Institute of Technology (Operations Research Center)

(contribution from Shashi Mittal)

Website: http://web.mit.edu/orc

Operations Research Center is an interdisciplinary program at MIT - it is not an independent department in itself. It has around 50 faculty members drawn from many departments, which includes EECS, Applied Math, Sloan School, Mechanical Engineering etc. The number of graduate students here is small - around 50 on an average. This makes the ORC a very closed knit group here. The OR program here is among the best in the world. Because of small faculty size, the faculty student ratio is generally good(i.e. not many PhD students advised by a single faculty member), though it depends from faculty to faculty.

OR is a multidisciplinary field, and encompasses several areas in Computer Science, applied math, management science etc. and is a highly application based field, applications being in diverse areas like computational finance, supply chain management, health care, and so on. Broadly, there are three main divisions of OR - optimization, stochastics and management science. The OR program at MIT has a good mix of all these three components in course work as well as in research. However, my feeling is that in recent years the research focus of the graduate students has shifted more towards management sciences (probably because one can earn big bucks in area of finance and consulting). Career options after doing a PhD in OR are academic jobs in business schools, applied math or operations research, positions in research labs(like IBM, Bell labs etc.), finance and consulting.

Why should one apply in Operations Research? OR is heavily math inclined. Apart from the usual theoretical CS stuff like discrete maths, algorithms etc. one also needs to have a good knowledge of probability theory and analysis for doing research in OR. Usually this is covered in the course work in OR programs, but one needs to have interest and aptitute for these things. If one is interested in applied aspects of theory then OR is a very good field go into because of its diverse applications.

Other good OR programs: Stanford Univeristy - operations research and management science(has good mix of CS and management science research), Cornell University-Operations research and industrial engineering, Berkeley(don't know much about OR there but most of the students here apped in Berk OR as well), Columbia University - Operations Research and industrial engineering, Princeton University (has a very good group in computational finance), CMU - Tepper School of Business.

MIT EECS - of course it is among the best EECS departments in the world. MIT usually has somewhat bizarre criteria for admissions. My feeling is that they tend to favor students of some particular institutes over others - based primarily on the performace of the grad students from that insitute at MIT (that's why lot of IITM junta here, but not very much from other IITs). MIT also looks at the kind of research projects you have been involved with, so good recommendations and publications count as much as CPI here. You have chance here if either you have an extraordinary academic track record (which means CPI of 10 or neararound that) or have a good research profile, i.e. publications in *top rated* conferences and journals.