bayeslord

Hey, I'm a final year undergrad at IIT Kanpur in Stats. I'm planning to apply for programs this year Undergrad Institution: IIT Kanpur (India) Major: Stats and Data Science GPA: 8.4/10 Type of Student: International GRE 170Q 169V TOEFL 113/120 Relevant courses: Statistical computing, stochastic processes (A), Statistical inference, Time series analysis, Regression analysis, reinforcement learning (B+) Analysis, Probability theory (both C+ unfortunately) Research experience: Worked for a few months on variational inference at the Uni of Warwick, publication unlikely. Working on learning and regret bounds for contextual bandits with an IITK prof, Work experience Quant research internship at a (not very famous) hedge fund Recommendations Two from professors at Warwick and one from a prof at IITK I did a couple of courses with. Not really sure about the strength Research interests Probabilistic machine learning, Bayesian non-parametrics(GPs, Bayesian optimization, active learning), statistical computing (MCMC, stochastic optimization), reinforcement learning My top choice PhD programs are Cal, Cornell, Columbia, NYU CDS but seems unlikely I'll get into any of these right now. As for why I want to apply to UBC- There are quite a few profs at UBC working in the areas I'm most interested in - Dr Geoff Pleiss, Dr Trevor Campbell, Dr Alexandre Bouchard-Côté, Dr Saifuddin Syed. Plus, research experience and a good GPA in a Masters degree could (?) compensate for my lacklustre undergrad GPA. Also, it's funded unlike most masters programs in the US. I'd love to have an idea about how competitive my profile is for the Stats MSc and I also wanted to know if there are any other good funded masters programs in Stats. McGill's seems good too but they don't seem to have UBC's depth in these areas in particular.

Thanks! Did you mean UCSC? I couldn't find many people in UCSB working on Bayesian stuff but UCSC seems really good!

I am trying to make my PhD apps list. The topics I'm most interested in are theory and applications to ML of: 1. Bayesian inference/statistical computing- MCMC and sampling methods, variational inference, normalizing flows and probabilistic ML (VAEs, diffusion models, etc) 2. Bayesian non parametrics esp Gaussian processes, (heard Austin is exceptional for this) 3. Reinforcement learning, bandit algorithms, Bayesian optimization (also related to second) Could y'all recommend good professors and programs in these areas (apart from the standard top statML ones CMU Cal Stanford maybe lol)? My profile for context: https://forum.thegradcafe.com/topic/167223-optimal-stopping-problem-for-phd-apps-is-a-year-for-an-ms-worth-it/

Only took real and complex analysis in my third semester, didn't do too well in those tbh (C+ i.e. 7/10). I'm unsure what I can do at this point to demonstrate my mathematical readiness- probably take graduate level measure theory and functional analysis courses in my MS and try getting good grades in those perhaps. (Also I missed mentioning that I did a Bayesian stats course as well)

Undergrad Institution: IIT Kanpur (India) Major: Stats and Data Science GPA: 8.4/10 Type of Student: International GRE 170Q 169V TOEFL 113/120 Relevant courses: Statistical computing, stochastic processes (A), Statistical inference, Time series analysis, Regression analysis, reinforcement learning (B+) Analysis, Probability theory (both C+ unfortunately) Research experience: Worked for a few months on variational inference at the Uni of Warwick, publication possible but probably after application deadlines Work experience Quant research internship at a (not very famous) hedge fund Recommendations Two from professors at Warwick and one from a prof at IITK I did a couple of courses with. Not really sure about the strength Programs I want to Apply for: PhDs in Statistics/DS. Also open to MS programs at reach schools if there's a decent PhD conversion rate Research interests Probabilistic machine learning, Bayesian non-parametrics(GPs, Bayesian optimization, active learning), statistical computing (MCMC, stochastic optimization), reinforcement learning Can someone suggest some good programs for my profile? My dilemma is whether I should apply this year or complete my BS-MS in another year. I think I should only apply this year to the unis I'd prefer over the expected outcome of waiting a year (probably something like Duke). My professor's advice is to not wait another year since it won't be worth the additional value to my profile in terms of chances of getting into reach schools that are meaningfully different in terms of opportunities and research quality. Benefits to completing an MS: * I would gain more research experience and possibly publish something * I could work with some of the professors at my uni who have realy great networks and get much stronger LoRs The downside is, of course an additional year. Would the potential upside in terms of P(getting into a reach uni like Stanford/Cal/CMU) be worth the extra year? Also, what unis would I stand a decent chance of gettin into this year?