5
$\begingroup$

As Yann LeCun mentioned, a number of PhD programs in data science will be popping up in the next few years.

NYU already have one, where Prof.LeCun is at right now.

A statistics or cs PhD in machine learning is probably more rigorous than a data science one. Is data science PhD for the less mathy people like myself?

Are these cash cow programs?

There is a huge industry demand for big data, but what is the academic value of these programs, as you probably can't be a professor or publish any paper.

$\endgroup$
  • $\begingroup$ This class of questions is being discussed on meta. You may voice your opinion here. $\endgroup$ – asheeshr Jun 11 '14 at 2:22
3
$\begingroup$

It seems to me that the premise of a PhD is to expand knowledge in some little slice of the world. Since a "data scientist" is by nature is somewhat of a jack-of-all-trades it does seem a little odd to me. A masters program seems much more appropriate.

What do you hope to gain from a PhD? If the rigor scares (or bores) you, then what about a more applied area? Signal processing, robotics, applied physics, operations research, etc.

$\endgroup$
  • $\begingroup$ As Andrew Gelman mentions HERE, it's pointless to be a math major if you are the best. I am not even that great at math, actually. That's what led me to think of data science. $\endgroup$ – user13985 Jun 10 '14 at 16:41
3
$\begingroup$

Computer Science is itself a multi-disciplinary field which has varying requirements among universities. For example, Stockholm University does not require any math above algebra for its CS programs (some courses may have higher requirements, but not often).

I am not sure what you mean by a machine learning program being more rigorous. They are just two different programs. Data Science would likely take a broader view and focus on application and management (business courses are maybe on offer?). The research could be rigorous in its own right, but it definitely won't be tailored to someone who wants to optimize new algorithms or solve the low-level problems of machine learning.

I don't see the Ph.D program listed yet in the link you provided. Will you please follow up here if you get more specific information?

$\endgroup$
  • $\begingroup$ You are right, the data science PhD is not yet listed. Machine learning has a lot math behind it, we can't just use an algorithm without proving the theory part why it converges, optimizes, etc... $\endgroup$ – user13985 Jun 10 '14 at 16:38
3
$\begingroup$

A cash cow program? No. PhD programs are never cash cows.

I don't know why you couldn't be a professor with a PhD in data science. Rarely does a professor of a given course have to have a specific degree in order to teach it.

As far as publishing goes, there are any number of related journals that would accept papers from somebody on topics that would be covered by the topic of Data Science.

When I went to college, MIS, Computer Engineering, and Computer Science were new subjects. Most of the people in my graduating class for Computer Science couldn't program anything significant at graduation. Within a few years, CS programs around the country matured significantly.

When you are part of a new program, sometimes it's possible to help define what it is that's required for graduation. Being a part of that puts you in rare company for that field.

As far as mathematical rigor is concerned, I would expect Data Science to leverage a heavy dose of mathematically based material. I wouldn't expect anything particularly new - statistics, calculus, etc. should have been covered in undergrad. Masters and PhD programs should be more about applying that knowledge and not so much about learning it.

$\endgroup$
  • $\begingroup$ You have very good points. But, imagine you are Data Science PhD grad, which journal will you publish in? Stat, CS, journal of machine learning? Each one of those fields has its own experts. $\endgroup$ – user13985 Jun 10 '14 at 19:31
2
$\begingroup$

No-one knows since no-one's completed one of these PhD programs yet! However, I would look at the syllabus and the teachers to base my decision. It all depends on what you want to do; industry or academia?

$\endgroup$
  • $\begingroup$ I would check out research by faculty and number of citations. Generally that is a good way to rank programs. $\endgroup$ – Stu Jun 9 '14 at 20:58
1
$\begingroup$

I think this question assumes a false premise. As a student at NYU, I only know of a Masters in Data Science. You linked to a page that confirms this.

It's hard to gauge the benefit of a program that doesn't exist yet.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.