James McInerney

2/2 Topics of interest from my side include probabilistic modeling & deep uncertainty quantification. We have a booth at #NeurIPS2023 this week if you want to chat with someone; I'll be there Wed & Thur late afternoon.

We are now open for applications for ML research internships at @NetflixResearch next summer. Great way to spend 3 months doing research, usually with aim of publishing. 1/2 jobs.netflix.com/jobs/300628646

I remember sitting on the 67th floor of the WTC with @allnoteson (in 2019?!), where he told me about his idea to build a dead-simple toolkit for AI developers. It's amazing to see how that idea has grown into @replicate. Congrats on the raise, excited for what's next!

Applications to @Mila_Quebec are now open. Come work with a colleague or myself! I work on projects in continual learning, probabilistic modeling, optimization, NLP, and a few selected applications (including recommender systems). Also looking to explore AI for education.

I am absolutely begging AI researchers to learn the data processing inequality.

The typical Bayesian NN setups treats all parameters stochastically. @MrinankSharma will shortly present our @aistats_conf "Notable Paper" on why this fundamental assumption is usually not justified and partially stochastic networks are often preferable proceedings.mlr.press/v206/sharma23a…

AABI is back! Co-located with ICML, in-person, and co-organized by yours truly :) Website: approximateinference.org #bayes #machinelearning #inference #aabi #symposium

@naivebaesian unfortunately not, we only take students enrolled in US PhD programs at this time

@mcinerneyj Do you sponsor visa ?

We are still looking for interns for ML research next summer. If you have a strong stats + ML background, I have a variety of exciting uncertainty quantification and time series modeling projects of interest. (Mark me as referral if you want me to take notice of your application)

@Karl_ma31415926 good question, the main internship program runs over 3 months in the summer and that's the period where you'd get the best experience. ad hoc arrangements are a special case I'd say.

@mcinerneyj Does Netflix allow spring/fall internship?

@pratham16181 thank you. I believe you can specify that at the next step if you're asked to fill out a form.

@mcinerneyj I applied, however, I did not see an option for referral.

this was a long research project, with many twists and turns, and an engaging collaborator in @nathankallus within a wonderful supportive environment at @NetflixResearch so grateful! poster: neurips.cc/virtual/2022/p… 6/6

it does this without ever having to explicitly compute the inverse Fisher info of the params (expensive or intractable in large models)! and both the infinitesimal and FD versions converge in the limit to the same result as the delta method (Thrm 1 & 2 in paper) 5/6

beyond excited to present our #NeurIPS2022 poster Thur on the implicit delta method, a surprisingly simple recipe for estimating epistemic uncertainty for evaluations based on deep models paper: arxiv.org/abs/2211.06457 poster: #327 Hall J, Thur 11am session w/ @nathankallus 1/6

@Aishwarya_R_M @dianarycai +1

@dianarycai I just got a latte at Revelator, it was pretty good!

Where are the good coffee places around #NeurIPS2022 ? ☕

@ltt_hal I'll be there! (same days)

Just booked my flight tickets and hotel to attend NeurIPS22 (28/11-02/12). Looking for partner in crimes for food tasting activities (and research discussions of course, but the stress is on the former). #NeurIPS2022 #gumbo #redbeanandrice

@psobot congrats! guide for "new" immigrants 😄

After almost six years in the U.S., four visas, hundreds of pages of paperwork, and a metric ton of anxiety: I can stay here now. 🇺🇸🗽