qqyyzz

I have considered that. Unfortunately the cs department here recently lost their main ML person, and there really aren't any machine learning people left at the cs department. In fact there is only one faculty that is working on anything AI related.

I'm currently in a stat PhD program and I just received my MS last spring from a top 20 ranked statistics program. I know that graduate students switching programs is relatively rare so I was hoping to get some feedback from experts on this forum on feasibility. If I were to switch over I would mainly interested in a PhD in CS focused on machine learning and data mining. Machine learning is studied in both computer science and statistics. Statistics takes more of a modelling approach to data while computer science takes a more algorithmic approach (for more details, see Leo Breimans paper: Statistical modelling a tale of two cultures). My main desire to switch is due to the cultural differences in how CS versus statistics treats machine learning. Many statistics departments have a very conservative culture. As an example, Andrew Gelman's blog mentions that many statistics departments have a measure theory requirement that is perhaps outdated. While it's useful for theorists he argues that for applied statisticians in a modern environment computing requirements would be much more useful. Additionally when it comes to machine learning, the statistics departments will generally be much more theory oriented. I'm not sure about how easy it would be to get into a CS program with my background. I come from mainly a math/econ background with a handful of CS courses. Here's a quick summary of my math, stat, CS courswork: Undergrad math coursework: calculus sequence, linear algebra, discrete math, ODE, numerical analysis sequence, advanced calculus (analysis), abstract algebra, probability, math stats. [(3.9 GPA for math courses)] Graduate Statistics coursework: theory sequence (casella and berger level), methods sequence (anova, GLM, boostrap, mcmc etc), design of experiments, stat computing, data mining, machine learning and multivariate statistics. [(3.8 GPA for stat coursework)] Computer Science coursework (mix of undergrad/grad courses): data structures, algorithms, pattern recognition (graduate), software engineering, algorithms (graduate - in progress), Artificial Intelligence (graduate - in progress). [(A's in all courses so far)]. In terms of research experience, I did have a research project for undergrad with a professor in the econ department on theoretical game theory (the project was funded by external grant if that matters). I've had a presentation at a conference and I also did a data mining internship this past summer. Unfortunately no publications (not unusual for stat as most students don't really start research until 3rd or 4th year). I think I have a fairly strong applied math/statistics background (with decent GRE scores, 90+ percentile in math/verbal) but I'm missing some traditional CS courses in programming languages, computer architecture, networking etc. From the research that I have done computer science programs seem to vary quite a bit in terms of admission requirements and PhD coursework requirement. What do you think are my chances of getting into a decent machine learning CS program?

Has anyone who received an offer from UCLA visited the department?

Look at the job placements for both: http://www.stat.ucdavis.edu/people/alumni-info and https://www.stat.osu.edu/people/alumni I think from the looks of it OSU has better academic placements.

Hi Everyone, I finally found out that I am waitlisted from UCLA. I know this is a lot to ask because people like to consider as many options as possible but if you guys have other offers and don't think you'll be going to UCLA please decline. I'm really worried as I don't have any offers yet. Thanks

What has happened with the program in the last 10-20 years that makes it overranked now?

Is that Ohio State or Oregon State? If it's Ohio State, I've visited their department and I think it's pretty solid in bayes and computing. Unlike a lot of other departments, I think they also offer promising students university fellowships so you might not have to TA/RA the first year. Also the Columbus area has a fairly low cost of living.

Thanks for the update. I guess my chances of getting into any programs now is pretty small

Anyone know what's going on with UC Davis? Will they send out more offers?

UC Davis has sent out acceptances. For those accepted, care to share your profiles? I haven't received anything and I'm starting to worry!

Is that UCLA stat or biostat? I don't remember UCLA stat giving interviews.

Anyone know what's going on with UCLA and Davis stat?

dasgut, I read one of your earlier posts and I think I remember you posting sometime in 2011 when I was also applying. Where did you apply the first time around? Your record looks better than mine when I applied.

I think as long as your lower GPA is in non-math classes it's fine. I was in the same boat as you except I had a much worse record coming out of undergrad. I was undergrad 3.3 GPA in economics from a liberal arts school. I worked at a bank for awhile, went back as a non-degree student and had close to a 4.0 in all math courses including many upper division and proof based courses and also worked as an RA in the economics department. I applied to a lot of schools and even with my abysmal undergrad record I was actually only rejected from a few schools (got into a lot and was waitlisted at even more). You can probably look at my post history and find out more details about where applied and where I got in.

A note about the UW-madison program. I got accepted to their program when I applied two years ago. I believe UW's biostat program doesn't accept students directly. You must apply and be accepted to the stat program. As a result if you're considering UW be warned that it will probably be more theoretical than a regular biostat program.

I'd be careful about UW (washington, not wisconsin) and UPenn as I know UW recommends the math GRE and UPenn "strongly recommends" the math GRE.

That's a fair point, I did have all three of those courses when I entered grad school, but at the time I submitted my application I only had real analysis and the first semester of advanced calculus was in progress.

Well that's good to hear. The only other mention of UCLA's stat program on this board is from a thread a few years old and the person did not have positive comments about their program.

Hi Cyberwulf, I always hear about UCLA's biostat program, do you know anything about their statistics program?

I think your application is pretty strong. I'm sure it will be stronger if you take more math courses of course but I think everyone here overvalues having graduate level analysis, measure theory etc on your application. I think part of the reason you didn't get accept to some those programs is your lack of GRE subject test, not necessarily due to the lack of graduate math courses. If you had read through their admissions sites, I know stanford and UW's programs clearly state they require the subject GRE. Uchicago, Columbia and I think Harvard "recommend" it. A lot of times something like the subject GRE is just a departmental requirement and could mean automatic rejection just not having it. However, I think not taking something like measure theory or other grad level math courses isn't an automatic rejection. When I was visiting schools a few years ago, I'd say about 80% of the students I talked to didn't have many graduate math courses and many never had measure theory prior to entering the program. The places I visited were not low ranked either, I'm talking about accepted students at places like cornell, wisconsin and UNC. In fact when I talked to the graduate director at UNC, he had said that I had more math background than the average accepted student and in terms of analysis I had only had 1 semester of real analysis (undergrad), and two semesters of advanced calculus (also undergrad). I was sort of in a similar situation when I was applying; I didn't take courses like number theory or topology or logic and as a result didn't have the required knowledge for the subject GRE. However, a lot of programs realize that the subject GRE is not a great measure for grad student success and a lot of programs don't care if you haven't taken it.

Right, a ton of people are saying fewer and fewer programs require it, but I have literally only seen like two or three programs where it isn't required. Where are these other programs that are causing people to say that fewer programs require it?

Cyberwulf, when I was applying last year not a single program had measure theory as a prereq for starting a program. I'm 100% sure Persi Diaconis is talking about "modern" departments do not require students to take measure theory. From reading UC Berkeley's program description it appears 215 (the measure theoretic probability sequence) is optional; there are other classes you can take as part of core. I think UCLA's program is the same way. Anyone know of any other programs like this? I'm curious which other programs are "modern".

I was reading an article on statistics education and there was a quote by Persi Diaconis about how there's a trend in statistics departments to no longer include measure theory as a requirement. Every department that I've looked at requires measure theory (with the exception of UC berkeley). I was wondering if you folks would know which departments no longer require measure theory for their PhD programs.

I was at visit day and biostats is certainly a large application area and their biostats department seems to have a ton of money. But overall their department is pretty large and you can easily find people working on other things. Also, it seems that a lot of departments work together fairly closely. So you can probably find co-advisors in computer science or other related fields. The one area where they're a bit underrepresented is economics side of things. So I think if you're interested in like econometrics and the like you're a bit out of luck. I also had a question about Purdue, did you get in? If so, what kind/how much funding do they get? When I emailed them yesterday they told me they were still looking at apps.

I know the decision deadline is soon approaching and I'm happy to say that I've been accepted with funding into some very good programs. In order to try to decide I've been looking at research interests and course listings. I am mainly interested in machine learning/data mining and I was hoping on getting people's opinion on this board. It seems a few of the programs I've been accepted to have a good number of professors that work on machine learning and data mining research. But in terms of coursework these departments for whatever reason do not have a dedicated course in either machine learning or data mining. I'm trying to decide between higher ranked universities that don't have coursework in ML/data mining versus lower ranked universities that have these courses. My main concern is that I would eventually want to do ML/data mining research or work in some applications that requirement ML/data mining. I'm not sure how difficult it would be to start doing research without some graduate courses first. I suppose I could take the courses from the CS department but that would require both a lot of prereqs from the CS department as well as worrying about which courses would count for credit. For the record here are some of the programs I'm considering: UNC, iowa state, wisconsin, ohio state and purdue. I've been visiting schools these two weeks as well.