cicada2014

Interesting. I can log in with both my user id and email. I doubt this means anything though. (we are an insane bunch, aren't we?)

Same. I really just applied to MIT for the hell of it and was not expecting to hear back. My interview was so-so. I fumbled over some questions but I think I did ok on others. Best of luck to both of us!!

same--had an informal chat with POI. But I am more of a borderline case so maybe they are chatting with borderline candidates. I wouldn't worry if you haven't heard back yet.

Ah ok. From the results section it seems like they released the fellowship awards first. Did you also get an interview at Berkeley?

It seems like USC CS decisions are out? Can anyone who has heard back post here (e.g. area of interest, stats, etc.)?

Yep. I actually talked to admissions faculty at a top school and he said that they pretty much don't look at GRE at all. The only red flag might be if you are an international student with exceedingly low verbal score. But even then, they give the benefit of the doubt and will actually try to talk to you in person to gauge your english ability.

GRE doesn't matter as long as your Q is above ~90th percentile. You're right at the cusp. I think you will be fine

Ok, make sure then your recommenders mention that the journals in which you published are tier 1. I still think you have a good shot at the top 4 (though not a sure thing of course). I think you would have a better shot at stats programs though. Look into schools that have faculty who are applying/researching ML in the context of biomedicine. Johns Hopkins, NYU, come to mind, but I'm sure there are more.

depends where your publications are i'd say. if they are in top conferences for ML (NIPS/ICML), then I think you'd be golden for all schools, including the top 4. I would knock Harvard stats off your list since they don't seem to be doing too much in ML. Look into (all CS) UCLA Columbia Georgia Tech Michigan USC

If you get the first author in ICRA and AI, then I am sure you will get in practically everywhere you apply. Actually even if you get one of those published as first author you will be competitive at all schools. If not, I think you would still have a decent shot at top 5-10 schools

Hi. I was wondering how the CS admissions committee views workshop publications? I've heard they don't really give that much weight to it. Reason being, my most recent paper was rejected from a top conference but accepted to a workshop (at the same conference). The workshop seems ok (and arguably more suited for the paper), but I think if I adjust a few things it may be accepted to another conference later in the year.

depends on what you want. If you got into CMU ML-stats joint program, then go there and don't look back (although Michigan's ML seems pretty good, CMU cannot be beat for ML).

i think if you have a tier-1 publication you will be a strong contender. I've heard that they give MUCH less weight to lesser conferences/workshops (of course, it's better than nothing). You are also at a disadvantage because your LoR will come from professors who are not likely to be known. Having said that, I would throw a couple of apps at the top 4 anyway.

yep, overall reputation does not matter for PhDs. go to CMU

why do you think you are interested in IEOR if you don't have any experience in it? if you are dead set on industry, then MS&E at stanford might make sense. otherwise, go to UW. UW applied math is very very good (don't be fooled by its (relatively) lower undergrad reputation. UW is a top graduate school for many STEM fields)

Hi, I'm trying to compile a list of schools to apply. I have a decent idea of which CS/Stats departments have strong ML faculty. I know there is some ML research going on at IEOR departments too, but was wondering if someone could help me compile a list? Thank you.

Then MSCS makes much more sense. She would be miserable in a PhD program if she doesn't want to do research

is it ok to cite papers in one's SoP? I want to motivate my research interests by citing some seminal papers in the area, but I don't want to come across as superficial. Is this kosher?

I was in a similar boat. Few options. 1) enroll in a master's program, preferably with a thesis component. Yes it will cost you an extra 30~40K, but this is the price you might have to pay for not pursuing research. 2) if you have a strong interest in an area, try doing some rudimentary research on your own. obviously you won't be publishing groundbreaking results, but if you get a publication out of personal research it will help somewhat. I was able to publish at a decent workshop through this method. 3) research position at a company that allows publishing. Google, Microsoft, Facebook etc. Obviously they are very competitive, but if you can get a software engineer position and worm your way into a research group, you will be set. Good luck!

Thanks all! I might apply to Stanford as well then, given that they are ridiculously strong at both NLP and ML. I looked into UMass, and know about McCallum, but I don't think their ML department is strong. I also looked into UIUC but wanted to live in a more urban area. But thanks all for the suggestions!

Hi all, was looking to get feedback on my list of schools (applying for Fall 2015). Brief background: Undergrad: Electrical Engineering from a top 10 school. GPA: 3.7 Grad: MS in Mathematics from a decent school (their math department is top 5). GPA: 3.8 GRE: 168/165 Q/V Publications: 1 publication at a mid-tier conference, 1 workshop publication (the workshop is of decent repute). Hopefully will have another one under my belt by the time I apply Research experience: This is where my red flag is. I don’t have any “official” research experience with a professor. The above publications were result of personal projects (with occasional guidance from professors) and/or class projects. My function at my current company is in an R&D role, but doubt that will help. LoR: Should be good-to-excellent, though probably not “top 5 students ever” category. Current List (my interest is in NLP and Machine Learning). Incidentally, these are the schools I identified as having both strong NLP and ML faculty. If I’ve missed any, please let me know (of course, the top 4 are good at both but I don’t think I have a realistic shot at Stanford, MIT, or CMU). Berkeley U Washington UCSD UCLA Columbia Upenn Brown USC NYU

please disregard. was supposed to post in CS forum

Thanks for the comments. Two additional questions: - How much attention do admission committees pay to the quality of publishing venue? What about workshops versus conferences? Reason being, I am preparing another paper and deciding whether I should shoot for a workshop publication at a top-tier conference, or a conference publication at a mid-tier conference. - I think my big weakness is that I don't have "real" academic research experience with a professor. My research has been limited to class projects, employment (we are barred from publishing), and personal interests. Will this hurt?

It seems pretty clear that the differentiating factor of applicants at the top school is one's research experience (pretty much everyone has has stellar GPA, GRE, good school etc.). My question is, how 'good' does your research experience need to be to be considered at the top 10 schools? My research experience has generally been limited to class projects, work experience (I work at a R&D group in my company, though it is more Stats-related than CS-related), and one personal project that I did in my spare time. Luckily, my personal project was accepted for publication at a decent (but not best) conference in my field. TL;DR: My research experience consists of one first author publication at a mid-tier conference, class projects, and work experience. Is this (combined with good GPA and recs) enough for me to attempt at top 10 (I know I am probably out of the running for the top 4).

Exactly. Much of the coursework assumed knowledge of undergraduate analysis, which put me in a lot of trouble. Oh well, can't change anything now I guess. At least I learned quite a bit, as we covered measure theory, Lebesgue integrals etc.