Profile Evaluation / PhD School Suggestions

[potentialphd2]

 

Background: Graduated in 2019 with double major in Math and Computer Science and then worked in finance industry for a year.  Returning this fall for a masters in the computer science department (at my school most statistics is under computer science) and will focus on some combination of machine learning / inference.  My areas of interest in statistics are bayesian inference, causal inference, theoretical machine learning, and theoretical statistics.  I am also interested in applications to healthcare.  I would like to evaluate my prospects for applying to PhD programs in the fall in Statistics (or Biostatistics) and Machine Learning, as well as to hear feedback.

Undergraduate Institution: Top 3 School.  I will be starting a Masters in Computer Science focusing on inference/statistics this year and will be applying to grad schools in the fall. 

Majors: Mathematics, Computer Science

Minor: Chinese

GPA: 3.6/4

Major GPA: 3.76/4

Type of Student: Domestic White Male

Courses taken:

• Math: Differential Equations (A), Linear Algebra (A), Real Analysis (no official grades that semester), Complex Analysis (B), Abstract Algebra (A), Intro to Stochastic Processes (A), Discrete Math Seminar (A), Logic Seminar (B), Intro to Probability (A)
• CS: Programming Fundamentals (A), Intro to Algorithms (A), Intro to Machine Learning (A), Algorithms II (B), Graduate Machine Learning (A), Graduate Seminar Machine Learning (A), Graduate Theoretical CompSci Seminar (A), Computer Systems (B), Software Construction (C)
• Stats: Graduate Mathematical Statistics (A), Graduate Inference (B), will also be taking 4 more courses in the grad statistics curriculum this year.
• Science: Biology I (No Official Grades that semester), E&M Physics (No Official grades that semester), Quantum Physics I (B), Quantum Physics II (C), Classical Mechanics (A), Intro Chemistry (C)

GRE General Test:  Not taken yet - seems like it might not be necessary? But happy to take if can improve my chances.

Research Experience: 

Spring 2015-Fall 2016: Machine learning research in the physics department at a different university applied to theoretical physics data, published in a physics journal.

Fall 2020: I will be starting a 1-year Masters this fall and will be doing statistics or ml-focused research.  

Working Experience:

Fall 2017: Teaching Assistant for Abstract Algebra I

Summer 2018, Sumer 2019-Summer 2020: Intern and quantitative trading analyst at bank

Letters of Recommendation: Could get one from past research advisor, one from current research advisor, past work supervisor, or academic advisor 

Currently considering schools:

PhD: What are my prospects at applying to top PhD programs (I.e. MIT, Stanford, Berkeley, UW, UChicago,...)?. 

Questions: 

Based on my background, am I better suited to apply to Statistics/BioStatistics or Machine Learning PhD Programs?

Do you generally apply to do research under a specific professor, or is this decided after acceptance?

Would it be more valuable to do masters research under:

1. A well-known quantitative finance professor (would likely do relatively applied Ml/stats related work on healthcare data)
2. A faculty in inference who is less well known (would probably do more theoretical work, the research group's interests are related to but not exactly bayesian/causal inference)

How big of a problem is it that I have no experience in my areas of interest outside of my classes?  Should I reevaluate?

I would love feedback from you all and I greatly appreciate your help!!  Please let me know if there's anything I can elaborate on!  Thank you so much!

[DanielWarlock]

I'm a starting phd myself this year so I can't claim authority on this. But here is my two cents. Normally I would be tempted to say that you will have no trouble at top programs as you listed. But I have seen too many competitive people here and elsewhere this year to say that you are a shoe-in (several published at places like anal of stats/prob, jmlr, jams etc). I think one of the reasons is that the covid pushed a lot of elite people to grad school from industry. An argument can be made that there is less competition from international applicants, but as far as I know most competitive people will apply regardless. But it is really hard to say. So even if your profile is very strong, I'd be cautionary and apply broadly: this does not mean you should not consider top 10 schools. But rather to apply to some range 20-50 schools as well. 

 

[bayessays]

1 hour ago, potentialphd2 said:

Based on my background, am I better suited to apply to Statistics/BioStatistics or Machine Learning PhD Programs?

Could definitely apply to any.

 

1 hour ago, potentialphd2 said:

Do you generally apply to do research under a specific professor, or is this decided after acceptance?

In CS, you apply to a lab. In statistics/biostat, you apply to department and choose an advisor after a year or two.

1 hour ago, potentialphd2 said:

How big of a problem is it that I have no experience in my areas of interest outside of my classes? 

Most statistics applicants don't have much statistics research experience at all. No problem at all.

 

1 hour ago, potentialphd2 said:

Would it be more valuable to do masters research under:

Doesn't matter, do what is most interesting to you so you can do well and get a good letter. 

I think you can pretty much apply anywhere for stat/biostat.  Pretty sure you'll get into top 5 biostat programs and top 20 stat programs. Don't just apply to the top 5 schools.  UW, Berkeley, Chicago are probably achievable but you can't guarantee with schools at that level. MIT/Stanford are probably reaches.  Schools like CMU, Duke, Michigan, and the top 3 biostat programs are probably good targets with a couple reaches and couple lower ranked programs. 

[statsnow]

In the top 10 no one is a guarantee.  It is all about fit. Lots of schools look at that closely.   Apply where you think you will the happiest. Some schools this fall are not using the GRE at all for admissions

[cyberwulf]

On 7/29/2020 at 6:01 PM, DanielWarlock said:

 Normally I would be tempted to say that you will have no trouble at top programs as you listed. But I have seen too many competitive people here and elsewhere this year to say that you are a shoe-in (several published at places like anal of stats/prob, jmlr, jams etc).

 

Give me one example of someone from this year who has published in Annals (Stats or Prob) before applying for a PhD.

[MathStat]

@cyberwulf I know someone who *submitted* a paper to one of the top 2 stat journals (Annals/JASA) (so not published but still extremely impressive). However, I assume it would be hard to have an actual publication before applying to the PhD since that would mean you would need to submit at least very early in your junior year (assuming you apply straight out of undergrad), which means you need to start the research at least sometime in your sophomore year. ---- Btw, how long does the review process for Annals of JASA take usually?

 

Edited August 1, 2020 by MathStat

[insert_name_here]

8 hours ago, cyberwulf said:

Give me one example of someone from this year who has published in Annals (Stats or Prob) before applying for a PhD.

https://web.stanford.edu/~chen96/index.html

[bayessays]

I saw someone with a JASA paper from their master's who didn't get into any top 30 programs. It's still rare.  Stanford statistics is really in a league of its own, so it's not surprising to me what an international student their had an Annals paper, but to be fair, cyberwulf did say "from this cycle" and this guy is a year into his PhD and the paper has still only been accepted and not published. 

[Stat Assistant Professor]

20 hours ago, cyberwulf said:

Give me one example of someone from this year who has published in Annals (Stats or Prob) before applying for a PhD.

I know a few Stanford alumni/students who published research that they had done as an undergrad (not their PhD research) in JRSS-B or Annals of Statistics -- often their work appeared in these venues during the first year in their PhD program. That is quite uncommon though, even for Stanford -- but it's not terribly surprising that a small number of folks at Stanford already have top-tier papers so early in their research careers, before they have even finished all their PhD coursework requirements.

I think it's more common to see international PhD students with papers in journals like "Statistics: A Journal of Theoretical and Applied Statistics," "Journal of Statistical Computation and Simulation," or "Journal of Business and Economic Statistics." I've seen this from a lot of Stat PhD students from South Korea.

Edited August 2, 2020 by Stat Assistant Professor

[statsnow]

I agree it’s very uncommon to see published research in a top journal before you start your PhD. Also with multiple authors sometimes it’s difficult to tell what the student contributed to the paper.