Fall 2020 Statistics PhD Application Evaluation

[ipsumlorem]

Hello,

I'm going to be a second year master's student in statistics in the upcoming Fall, and I want to apply to PhD's in statistics. At this point, I think I want to go into academia with a focus on machine/deep learning. Some concerns with my application is that my math background is not the greatest, and my master's GPA is a little bit lower than it should be. Any feedback would be greatly appreciated!

Undergrad Institution: Top 30 USNews Public
Major(s): Statistics, Economics
Minor(s): Hispanic Literature and Cultures
GPA: 3.79 (Major GPA 3.82)

 

Graduate Institution: Stanford

GPA: 3.87

Type of Student: Domestic Asian Male

 

GRE General Test: 169Q/160V/5.5AWA (TBD on Math subject test, but I'm not too confident in it)

 

Research Experience: I have an Economics undergraduate thesis that dealt with NLP and econometric modeling. I'm doing research at Stanford working on convex optimization. My current internship deals with deep learning related to MRI data.
Awards/Honors/Recognitions: Undergraduate highest honors in economics
Pertinent Activities or Jobs: I currently do some analytics for the women's basketball team in season. My current internship, as said before, deals with deep learning related to healthcare.

Letters of Recommendation: One undergraduate economics professor, the professor I'm doing research under at Stanford, and my supervisor at my current internship. They should be good
Undergraduate Courses: Probability (A), Linear Algebra (B), Statistical Theory/Inference (B), Calc 3 (A), Discrete Math (A-), Econometrics (A-), Stochastic Processes (A), Time Series (A), Diff EQ (A), R Programming (A), Object Oriented Programming (A), Optimization (A)

Graduate Courses: Real Analysis (A-), Statistical Inference (A-), Machine Learning (A), Theoretical Bayesian class (A+), Applied Bayesian class (A-), Measure-theoretic Stochastic Processes (B+), Advanced R class (A), Convolutional Neural Networks for Computer Vision (A), Regression Modeling (A)

 

I also plan on taking advanced regression modeling and proof-based linear algebra this upcoming fall.

 

Program: PhD in Statistics

 

School List: I have no idea how my profile stacks up due to my lack of math background, and I'm not sure which schools to apply to. I also feel my current list is way too top-heavy.

 

Stanford

Berkeley

Harvard

Washington

CMU

PSU

Columbia

OSU

UCLA

UT Austin

UC Davis

UCSB

Maybe Cornell

 

Any feedback is much appreciated

[bayessays]

I'd say the top half of your list is reaches, some of them being extremely unlikely (unless your Stanford professors love you and want you to stay). The bottom half of your list looks like reasonable targets.  I'd add another couple mid-range schools like OSU (think Minnesota, UIUC) and a safer option or two.

An issue you're going to have is finding deep learning faculty in stats departments.  I know UT Austin has a couple faculty doing it and the top departments will as well, but in the schools that are in your sweet spot they are few and far between.

[ipsumlorem]

14 hours ago, bayessays said:

I'd say the top half of your list is reaches, some of them being extremely unlikely (unless your Stanford professors love you and want you to stay). The bottom half of your list looks like reasonable targets.  I'd add another couple mid-range schools like OSU (think Minnesota, UIUC) and a safer option or two.

An issue you're going to have is finding deep learning faculty in stats departments.  I know UT Austin has a couple faculty doing it and the top departments will as well, but in the schools that are in your sweet spot they are few and far between.

Thanks for the response! Is it safe to say that I should stay mostly in the 10-25 range of USNews schools?

And yes, I've realized that deep learning is not very prevalent in the mid-tier schools, so I've been having some trouble with that.

I've also been looking at the PhD in Machine Learning, but my CS background is good but not great and I wasn't sure about how that would translate over in getting into academia. I mainly want to be a professor and teach with some research on the side as a goal.

Edited July 21, 2019 by ipsumlorem

[Lp_space]

Your math gpa may be too low for top schools. Given your B's in linear algebra, math stat and measure-theoretic stochastic processes, the admissions committee may doubt your ability to handle PhD level courses. Also, you have very few proof based courses compared to other applicants to top schools, which doesn't help your case. With that being said, you should look into schools in the range of 25-40, and maybe add a few more safeties. I would identify schools at the level of OSU and Davis be your reasonable targets.

[AnnieM]

I wouldn't sell yourself too short, you never know, it's such a crap shoot and why not aim high while keeping a decent safety net?  I was a Biostat Masters grad from UC Berkeley, and spent a lot of my time working in the Berkeley Stat department - and what I learned from their grad students as well as their policies is that they value high grades above all else.  The masters students in the Stat department were automatically not offered spots in the PhD program after their graduation if they earned even one B+ in their masters coursework.  The grades had to be all A's.  So, it may be tough for you there, but you never know.  It's worth finding out what each department values.  I honestly think you seem very well-prepared for a machine learning PhD and certainly a great program exists that will be an awesome fit.  Stanford, Berkeley, Harvard, Washington - those are the top-of-the-top schools, basically a long shot for nearly everyone.  Doesn't mean you shouldn't apply.  I visited UCLA (for Biostat, not Stat), and I loved the vibe there.  I also knew several grad students in the UC Berkeley Statistics dept that came from UC Davis Stats dept.  Have you considered NC State and/or UNC?  Duke?  What about Michigan, Wisconsin-Madison, or Minnesota-Twin Cities?  All have great programs from what I hear. Best of luck