University suggestion for Machine Learning?

[Desi_Mama]

Hello everyone,

 

I was wondering if someone here could suggest strong universities in the area of machine learning. This will help all future applicants as well.

 

To put my background in context, I did not do by BS in computer science, but rather in mechanical engineering. A better description of my coursework could be "operations research" since all my courses and projects were heavy on the optimization and networks side. I found the area of machine learning and developed my interest in the same at a late stage, so wasn't able to change majors to CS or to take courses in CS to cover breadth and apply for PhD programs in CS.

 

I did manage to get into a few mid ranked PhD programs and the MSCS program at Stanford. After some deliberation, I have decided not to take up the latter (too costly). The PhD programs are not particularly to my liking, and I feel I can do better with one additional year of CS related research. I will be joining a very well reputed R&D lab (along lines of MSR, IBM etc.) as a research assistant (intern) for this year for a project related to convex optimization and machine learning. I am planning to either defer my PhD offers and apply to other programs, or decline the PhD offers and simply apply again this December.

 

Can you suggest some universities which are very strong in machine learning? I am more inclined towards the math part (closer to optimization) as opposed to applications, though I would be OK with that too. I fear the top programs like Stanford, UCB, and CMU might be out of my reach (I will try none the less), but are there any other universities particularly strong in ML? Please let me know.

 

Also, if someone is aware, can you suggest some Operations Research programs strong in the area of Machine Learning. So far, I have been able to pin down only Columbia University.

 

Following is based on my research thus far:

• Tier 1: Stanford, CMU, MIT, UC Berkeley
• Tier 2: U Washington, UIUC, Univ. British Columbia etc.

It would be great if more people can help fill the list. Thanks

[phdreally]

You could join the PhD program you are currently admitted into (anything in the top-20 are generally worth pursuing), and then try for a transfer the next year, if you really want to. You might be able to carry forward credits earned in the first year.

[phdreally]

Also, look for professors, and apply to the universities where they are currently homed. The professor is your goal, and their university admission is just a means to that end.

[Desi_Mama]

Also, look for professors, and apply to the universities where they are currently homed. The professor is your goal, and their university admission is just a means to that end.

Thanks for the inputs. Given the circumstances, the intern is better than going to a university and try to transfer. That introduces too many new variables and a lot of anxiety. Also, the intern is reasonably well paid, and I get to work with some scientists whose research profile can match many professors in top 20 universities. Also, I can get a recommendation letter from them, unlike in the case of transferring.

Also, I feel that a major reason for why I got rejected by many programs (I applied to 12, got into 2 PhDs and 1 MS ) is no recommendation letters from people with PhD in CS. All my recommenders either had a degree in MechE (2 of them) or Industrial Engineering (the third). I can maybe use this year to rectify some deficiency too - do a discrete mathematics and operating systems course online etc. So I am pretty set on the intern position.

I am either looking to defer offers and apply to a few more programs (this would be ideal), or just drop the offers and apply afresh (this scares me a bit, but if it comes to that, I have no other option I guess).