Itay Evron

Why does L1 regularization induce sparse models? Many illustrate this using the least squares problem with a norm constraint. The least squares level sets are drawn next to the different unit "circles". I prepared a cool animation which I believe makes it even clearer 🙂

@Theo_Vincent_ It happens sometimes even without LLM involvement

While reviewing for ICML/RLC, I was surprised to find submissions citing the same work twice with slightly different BibTeXs🤔 It is always nice to see authors giving double credit to the works they rely on!👌 ...or it is an artifact of an LLM helper😅

How does an LLM’s past influence its future?🤔 In our new paper with @FazlBarez,@mtutek,@boknilev, Shay Cohen, we show that conversational history creates a "geometric trap" in the latent space, confining the model’s trajectory➡️making old habits e.g. hallucinations hard to break

🚨 2026 @Princeton ML Theory Summer School Mini-courses by: - Subhabrata Sen @subhabratasen90 - Lenaic Chizat @LenaicChizat - Sinho Chewi - Elliot Paquette @poseypaquet - Elad Hazan @HazanPrinceton - Surya Ganguli @SuryaGanguli August 3 - 14, 2026 Apply by March 31. Link 👇 Sponsors: @NSF, @PrincetonAInews, @EPrinceton @JaneStreetGroup, @DARPA, @PrincetonPLI, Princeton NAM, Princeton AI2, Princeton PACM Some amazing speakers from previous years: @Andrea__M, @TheodorMisiakie, @KrzakalaF, @_brloureiro, @rakhlin, @DimaKrotov, @CPehlevan, @SoledadVillar5, @SebastienBubeck, @tengyuma

@tsoofbaror x.com/itayevron/stat…

כבר הרבה זמן שהמטריקה היחידה שאני עוקב אחריה באדיקות היא זו. המודלים הופכים לחכמים יותר, זה נכון - ובמקביל לחרטטנים יותר.

Please rewrite your proof, fully rigorously. - I can’t honestly rewrite that previous document into a fully rigorous proof because the core mechanism I relied on does not hold in the way the draft claimed.

1/6 🧵 Calibration is hard. Multicalibration—fixing errors across every possible subgroup—is usually impossible at scale. Until now. Introducing MCGrad: A production-ready multicalibration library from Meta, accepted at KDD 2026. 🚀 github.com/facebookincuba…

A glimpse into the research I’ve been leading over the past year at Meta 🥹 . So many organizations own rich graphs that remain largely underutilized. GraphBFF shows how to build feasible, powerful Graph Foundation Models from these graphs, end to end, from data curation and modeling choices to production. We rely on real data, and solve real problems, no toy setups, just what it actually takes to make a Graph Foundation Model work in practice. This has been a life-changing experience for me, taking something from an idea all the way to a deployed GFM that is now having real impact at Meta. The preprint is now available on arxiv.

@miniapeur Higher-Order Learning Dynamics on Cellular Complexes M Alain, Terrence Tao, Yoshua Bengio Annals of Mathematics, 2026

@itayevron x.com/miniapeur/stat…

An LLM hallucinating a citation with your name on it is just the universe telling you to write that paper

📌 [1/4] A Graph Meta-Network for Learning on Kolmogorov-Arnold Networks We introduce a weight-space model for KANs, where learning happens directly over the KANs' 1D functions. This work was done during my Meta internship. openreview.net/pdf?id=ONpyYav…

@boknilev @HebAcademy ועל כך כתוב בספרי זוטא: תמהני עליך יונתן שאמרת דבר זה

@HebAcademy תל־אביבה

לירושלים – ירושלימה. לתל אביב – ?

@aryehazan Are you a reciprocal PC now?

I'm submitting 400 icml papers wish me luck

@OmarRivasplata @bremen79 And a sleep command

@bremen79 With a for loop?

ICLR decisions are being rolled out by ID number

@tsoofbaror 🎯

הAI יחליף את כולנו? לא יודע מה איתכם, ככל שהAI ״מחליף אותי״ יותר אני עובד יותר שעות.

We are back. For the first presentation of the year, Uri will present his recent work on agnostic reinforcement learning. See you next Tuesday!

(This is, of course, a simplified statement which refers to the *worst case*)

One of my papers I'm especially fond of, now accepted to ALT2026. 🥳 A question kept me busy for a few years: Do continual linear models under random task orderings converge more slowly in high dimension? By reducing this problem to stepwise-optimal SGD, we show they do not! pic.twitter.com/iiFzqlHb1M

In memory of my father, an educator who loved the Mishna. One teaching he cherished describes four kinds of students sitting before the sages as "tools". Adapted to modern terms, these are: A sponge, a funnel, a sifter, and a colander.

Remember our ICML25 "Graph Learning Will Lose Relevance Due To Poor Benchmarks"? Fear no more! GraphBench is here! 🤩 We give you: The next generation of Graph Benchmarking! Including: -New shiny high-quality datasets from diverse domains spanning seven domains, including chip design, algorithmic reasoning, and weather forecasting. -Standardized hyperparameter tuning procedures, enabling fair and principled model comparison - Strong, transparent baselines that accurately reflect algorithmic progress - Comprehensive coverage of graph learning tasks, datasets, and modern GNN architectures - Reproducibility-focused design, minimizing variance and evaluation artifacts - Forward-looking benchmark designed for next-generation graph learning research A huge collab with: @chrsmrrs, @mmbronstein, @michael_galkin, @HolgerHoo, Timo Stoll, @ChendiQian, @benfinkelshtein, Ali Parvis, Darius Weber, @ffabffrasca, @HadarShavit, @antoinesrdin, Arman Mielke, Marie Anastacio, Erik Müller,

Since linear probes are popular again, maybe it’s a good time to point to the many issues with them, which were examined in detail in the NLP Interpretability community. The “mechanistic?” piece by @sarahwiegreffe and @nsaphra has many useful pointers. aclanthology.org/2024.blackboxn…