Area specific courses for an MS/PhD degree for a non-CS undergrad

[ev a.]

Hello, 

 

I am an Electrical Engineering undergraduate student with following background:

 

->VHDL, C++, C, Java, Assembly (8051), PHP, javaScript, MATLAB like languages...

->Many courses of CS undergrad curriculum like linear algebra, discrete math, probability theory, logic design, ... overlap with my original curriculum.

->I have taken Algorithms and Data Structures courses.

 

 

I want to study in CS, specifically  in Computer Vision and Machine learning area. For this purposes, I am planning on taking additional courses like Machine Learning, Image Analysis (computer vision), Parallel programming but not Programming Languages and/or Computer Architecture (somewhat similar to a course I have taken that is related with Microprocessors) or Operating Systems, Database Systems and so on.

 

QUESTION: Is this advisable? (Taking area specific courses for a Masters application instead of filling the undergrad fundamentals) (Would I increase my chances for an admission from EE to CS with this strategy?)

 

I am planning for a PhD if possible but I am not sure the universities (BY THE WAY: I am planning on applying to TOP100 schools in CS subject only) will be happy with my somewhat detailed area specific approach compared to the other one?

 

 

Note: I know that there are zillion factors for admission but lets assume everything is static and this variable is the only determinant, would it increase my chances then?

Thank you for reading this long post!

[shinigamiasuka]

Have you checked the ECE/EE departments of potentials schools you're interested in? I'm asking this because some departments (like UMich's EECS) have computer vision and machine learning within EE/ECE rather than in CS/CSE/CE, usually as a part or an extension of the signal and image processing area. Even if they are within CS/CSE/CE, it wouldn't be problem for your graduate degree since you'd be studying a specific area.

 

In addition, the courses on computer vision and machine learning are either senior or graduate level courses anyway, so you don't have to worry much about not having the courses done yet. If you do get about taking those courses though, better for you and your chances for a PhD in CV/ML, just that you wouldn't need to. For a PhD, you'd need research experience and/or publications. So if you can get some (related) research with a professor in CV/ML at your school done right now, that's what's going to count. 

 

So yeah, you're fine applying whether it be in EE/ECE or CS/CSE/CE.

 

But more importantly, are you ready to jump into a PhD right now? You could possibly talk this out with a professor or an advisor, especially in CV/ML. If you're a junior or a rising senior asking this, then you still have time left to finish your undergrad. In case you're still not sure after talking it out, I'd advise you to take the courses, see how you like them, and in the meantime join a professor's research in the area. Then, you'd have a much stronger case for yourself.

Alternatively, you could take the MS route, which could either be very expensive or not depending on whether or not you get funded. It's much easier to get admitted into MS, and you wouldn't need much of a background in terms of research.

[svent]

CS departments want to see you've taken Theory of Computation, OS, Architecture, and most likely Programming Languages. You should take these courses. This would be like applying to a Math Ph.D. program after taking algebra, algebraic topology, algebraic geometry, p-adic analysis, and representation theory, but never taking a real analysis course.

[ev a.]

@shinigamiasuka

Thanks for the input!i'm lately starting to think ms would be better compared to phd since i'll have better chances at a higher ranking school after a productive ms ed.

CS departments want to see you've taken Theory of Computation, OS, Architecture, and most likely Programming Languages. You should take these courses. This would be like applying to a Math Ph.D. program after taking algebra, algebraic topology, algebraic geometry, p-adic analysis, and representation theory, but never taking a real analysis course.

Hello svent,

this is a very specific question but bear with me. I want to specialize in CV & ML yet you seemed to imply that departments don't care about above graduate level topics like ML as much as much as they do for the fundamental knowledge.

So for a top 100 school, is taking a Pattern Recognition course less preferable to Prog. Lang, course? I actually made a list of courses I can take and since I don't satisfy prerequisites of CS courses like Computer Architecture and Operating Systems, I went on an opposite route, made a program of Parallel Programming, an undergrad CV course, a grad ML course (i'm still an undergrad) and finally a PL course. I have also two other courses for EE I should take. Is this not advisable? Would I be better off just sticking with as much as undergrad courses I can take? But there is the problem of even I want to, I can't take some of the courses immediately since I have to fulfill prereqs of 2-3 courses to reach to that point. (and unfortunately i'm in my last semester)

[svent]

Do what you can. Showing specific interests is important, but know that ignoring pre-requisites for admission is risky. There are other fish in the sea.