Jump to content

CauchyProcess

Members
  • Posts

    57
  • Joined

  • Last visited

Profile Information

  • Gender
    Not Telling
  • Application Season
    Already Attending
  • Program
    Statistics & Applied Mathematics, PhD

Recent Profile Visitors

1,699 profile views

CauchyProcess's Achievements

Caffeinated

Caffeinated (3/10)

0

Reputation

  1. How close to complete does a particular research topic have to be for it to be appropriate to post a preprint on arXiv? I'm in the following situation. I've developed a new computational algorithm for solving certain problems that is completely parallelizable, but do not yet have a proof of convergence. I have strong numerical evidence of convergence, and very good theoretical justifications to believe that the algorithm should always converge, but writing down an actual proof is highly nontrivial - most likely to the point that me and my advisor would need to collaborate with a specialist (likely a pure math professor) in order to work everything out. My advisor is asking me to write up the algorithm, together with practical results on how to initialize it, tune it, etc., and post it on arXiv so that people can start looking at it and/or using it, but personally, I'm hesitant - it just seems early, premature. What is the norm regarding preprints? Is it appropriate to write up a short document describing the algorithm and post it on arXiv even without a proof that said algorithm works? Is arXiv the correct place, or is this better suited for something like a tech report?
  2. I'm going to soon start year 2 of my graduate program, and have one final opportunity to apply for the NSF fellowship. So, I'm wondering: can the research proposal be current research (with partial results)? I've developed a novel computational algorithm to be used in statistics that is completely parallelizable, but don't yet have a proof of convergence. I'd like to use this algorithm as my research proposal, and include a description of numerical results that I've obtained on large data sets. The algorithm works as far as I can tell: it finds the correct answer for simulated data, and finds the same answer as other techniques on real-world data. The idea is new, and entirely my own (i.e. not a project from my advisor). Is this appropriate - or should the NSF proposal be an area of research that is novel and that no-one has yet worked on?
  3. Hello everybody, I am starting grad school very soon, and wanted to write a short post asking for any last-minute advice for my next 5 years. Here's my situation: I'm an incoming PhD student in statistics. I'm attending a relatively new department/program, with all of its research very narrowly focused in one area. The program is unranked due to its young age, but has some well-known faculty. I am attending on full fellowship, for 5 years. My main goal within the program is to learn: getting a bachelor's in statistics just did not teach me very much on the subject. My career goals are fairly open: I'd love to work in academia, but feel that it's likely I will be boxed out of such a career due to the non "top xxxx" nature of my university, no matter what the quality of my own research is, so I'm keeping my eyes open toward industry careers as well. I've been spending my summer working: analyzing data for a healthcare company, they're interested in maintaining a relationship with me long-term and want to look into consulting services with the department that I am joining as well. Due to summer work, I am in good shape financially, but will still have a good deal of student debt while a grad student. I have also accumulated a few research ideas stemming from my previous undergraduate research at an NSF REU, and previous NSF fellowship application, but I don't know whether I will have the mathematical preparation to begin anything anytime soon. I am moving up within two weeks, and will reside on campus for my first year. Anything I should do? Think about? Any misc advice you would give me? I am open to any and all ideas.
  4. One of the difficulties with mathematical research is that it can be unbelievably hard to explain to others. I've spent a bit of time studying measure theory, and I'm sure your friend's research is interesting - but probably way over everyone's head, and certainly not explainable in a few days worth of time. Something about math also seems to attract people that can be awkward at times. Being very social myself, I've never understood it. edit: typo
  5. 1. Attending: UC Santa Cruz 2. Turned down: SUNY Stony Brook, UC Irvine, UCSB (my undergrad) 3. Reasons for attending: got really excited about Bayesian statistics after reading more about the field and $24k full fellowship is a much better financial offer than the TAships the other schools offered. Some random bits for future PhD applicants: - Don't forget basic conditional probability. Your odds of admission at similarly-ranked universities are highly dependent, and thus you are likely to either get into all, or get into none. Therefore, don't waste money applying to 8 top-ranked schools, of which you have 2 favorites: if you get into one of the 6 non-favorites, you'll probably get into one of the 2 favorites as well. - Take the GRE seriously. Admissions committees don't care about it, but if you can get 95%+ on all three categories, your odds of getting a university-wide fellowship are far greater than if you blow off the writing section and get like 37%. - Your undergraduate institution matters a lot. Unless you are absolutely exceptional (i.e. took two full years of graduate stats courses as an undergrad and ready to pass your quals the day you start your PhD), you may get boxed out of top schools, even if you have perfect grades/good recs/etc, just based on not being in a top-20-math-type school. - The applicant pool depth seems to be growing exponentially. Your year will be far more competitive than ours. - Remember, the math does not change based on institution. You will probably be learning from the same textbook whether you are at Stanford or UC Riverside. - Good luck!
  6. Didn't get it, didn't get HM. G/G VG/VG VG/E
  7. We're already up to page 54 on this thread. Last year, the results were released by page 26. Based on this, I have a feeling competition is going to be much stronger this year.
  8. Alright folks, I'm off to watch the 30minute weekly long-term weather forecast. Hopefully the results are up by the time it's done.
  9. Based on previous thread, was up at like 3:44am or something like that last year.
  10. I feel like the maintenance isn't going to be completed exactly on time. It's either going to be earlier or later.
  11. Here's a small update to anyone that may be interested. Applying to Where: Stanford - Management Science & Engineering, Probability & Stochastic Systems Track / Rejected from PhD, accepted to Master's with no funding, can't go Columbia - Statistics / Still pending as of March 28th, presuming rejection Princeton - Operations Research & Financial Engineering / Rejected Brown - Applied Mathematics / Rejected UC Berkeley - Statistics / Rejected UCLA - Statistics / Waitlisted NYU Stern - Statistics / Rejected UCSC - Statistics & Applied Mathematics / Accepted, 24k fellowship UCI - Statistics / Accepted, 17k TAship UCSB - Statistics / Accepted, no info on funding SUNY Stony Brook - Applied Mathematics & Statistics / Accepted, 24k TAship Just finished visiting all of the departments that accepted me, still need to think a bit about what I want to do, but leaning UCSC since it seems the more I've been reading papers/books/etc about Bayesian statistics, the more interested I've gotten in the topic. And, they offered me a full fellowship, which is big plus because I don't want to TA. That's no doubt the case for me given my posts, and precisely the reason I did not post anything until a good deal into the process, at which point my applications would have probably been read already. But in the end, this is just some forum on the internet - are faculty really going to be spending their highly valuable time reading it?
  12. Yes. Have not heard a word, but they are a government agency and very strict on procedure, usually don't publish results until early April, and publish all at once - rejections, acceptances, etc for everyone.
  13. Thanks for article, Applied Math to Stat! Their ideas on looking up NSF grants, citations, and the like seem to be quite useful, and I'm going to collect some data and see what I find. It's clear to me that once a program has a ranking/reputation, it won't change very much if at all over time. What's not clear is how that is initially established once a department is first created. For example, UC Irvine's Statistics department was recently founded by faculty that went there from UC Davis. UC Davis has a fairly well-regarded program, does that mean that UC Irvine's program will also be fairly well-regarded over time? Or, will it start at the very bottom of the rankings, and slowly move up? I would also hope that a committee would pick the applicant with stronger publications, but I am not so sure that in practice this would happen, given that academia is a fairly small world and knowing other people is very important. It's really hard to tell how it all works while being a mere undergraduate, so I'm inclined to not think about it too much now, but rather decide later on whether or not academia is a good fit for me, especially if I face an uphill road and have better options elsewhere.
  14. I have two admissions offers from relatively new (<10yr old) departments at UC Santa Cruz and UC Irvine. For example, UC Irvine has little placement data, as they only currently have 4 PhD graduates. What route would you all recommend taking in evaluating the strengths/weaknesses of such programs? Over the long term, what sort of factors influence the reputation such programs develop?
×
×
  • Create New...

Important Information

This website uses cookies to ensure you get the best experience on our website. See our Privacy Policy and Terms of Use