Jeffrey Negrea

New preprint! This is a hardcore technical paper on Thompson sampling - as a strategy for the so-called online learning game. I think it's one of the most long-term important things I have ever worked on due to what it makes possible. But that needs explaining - thread below!

I’m excited to announce that in July 2025 I will be joining @UWaterloo as an Assistant Professor in the Department of Statistics and Actuarial Science! Until then, I will continue at Princeton as a DataX Postdoc Fellow, working with Boris Hanin. I have many exciting projects lined up to study the scaling limits of neural networks and their training dynamics, with some important practical implications on the horizon. Stay tuned!

@roydanroy @ArnaudDoucet1 @wellingmax I think the analogy needed to make ``optimal scaling'' work is that each of your papers is an infinite product of block-iid sections.

@ArnaudDoucet1 @wellingmax Block IID, suffices, no? @jeffNegrea

I have noticed that the “review process” is very similar to mcmc: the further away from the current state you propose the less likely it is to be accepted”.

Congratulations to Dr @HaghifamMahdi on a successful thesis defense at U of T! 🌟🔥🎉👏🔥🌟🎉 Next stop: Distinguished Postdoctoral Fellow at the Khoury College of Computer Sciences at @NorthEastern.

3.5 years after we started this project while visiting @the_IAS, it’s been accepted to the Annals of Statistics. I won’t be at #JSM2023 but make sure to stop by Jeff’s session to hear about our work!

Super excited that this work has been accepted for publication in AoS. For those at #JSM2023, I'll be speaking about it in the session on Advances in Statistical Machine Learning, scheduled for Sunday 8/6/2023 2:00:00 PM. I'm speaking first so don't be late 😉.

Congrats to @jeffNegrea and @blairbilodeau. Paper here: arxiv.org/abs/2007.06552

Annals: ✅

I'm _incredibly_ excited about this paper. Deep insights for prompting, and for how models represent latent knowledge about the world. See the paper and check out @wzihao12's thread! Some extra (more speculative) nuggets below:

Prompt engineering is a dark art. To understand limitations, we formalize "what the user intended" through latent concepts. Then we show direct model controls by algebraic operations on suitably chosen representations!github.com/zihao12/concep… w @ybnbxb @jeffNegrea @victorveitch

@miniapeur Who needs a VC? Just "pull yourself up by your bootstraps"!

A concept for free energy. Any VC around?

@PhDemetri @finbarrtimbers something something purrr-ist

@finbarrtimbers There is always a small sect of purists. They get smaller each year.

It’s actually not unusual. Like at all. It would be weird if you learned R by reading those message boards, which are notoriously unwelcoming.

@kareem_carr Only the avatar can master all 5(?) elements

no data scientist has mastered all 5: 1. R 2. Python 3. javascript 4. making eye contact 5. Microsoft Excel

Congratulations to Dr. Jeffrey Negrea (@jeffNegrea) on the successful defense of his dissertation! No corrections. Superb job.

@kareem_carr Sort columns separately first

Make a data scientist cry with just 4 words

@roydanroy @Usman_skhan Influence by osmosis => coffee was spilled on them

@Usman_skhan All of them crossed my desk at some point. Many of them sat there for quite some time and I'm sure that there was some influence through osmosis.

135 papers submitted, a record! Congratulations to me. Thanks in advance to all of you who will be reviewing these during your summer. Many are just undergrad ML course projects that I was too embarrassed to kill earlier. Sorry not sorry 🙇🤦‍♂️🪺

@roydanroy it expects labels to be all caps sometimes, and it seems safer to wrap \crefs in {} within titles. doing that and defining the label twice, once with all caps and works for me... dm me a mwe if not

@jeffNegrea I've found that this package can sometimes not play well with other packages. Indeed, it barf'd in overleaf in a neurips paper. Not sure how close I was to fixing the issue.

Token not allowed in PDF string? You're probably including math or a reference hyperlink in a section title. Here's a hack for references. \newcommand{\saferef}[2]{% #1~\texorpdfstring{\ref{#2}}{\ref*{#2}}} as in \section{Proof of \saferef{Theorem}{thm:mainresult}}

@trongthangpham @ccanonne_ I assume they are custom made using tikz or something. Not certain though.

@jeffNegrea @ccanonne_ May I ask how do they have those notations? Are those from a font or something?

@ccanonne_ (but actually this is really well thought out notation that just looks confusing at first glance)

@ccanonne_ numerical analysis: hold my beer