Turns out pointmaps are not required after all and you can train rotation prediction via L1 loss on quaternions.

New #Featured Certification: DINOv3 Oriane Siméoni, Huy V. Vo, Maximilian Seitzer et al. https://openreview.net/forum?id=2NlGyqNjns #visual #features #supervised

#illusion le damier qui enfle / bulging checkerboard. Wearing polka dot clothes makes you look larger. Porter des vêtements à pois fait grossir

Accepted to ICML 2026!

Accepted to ICML 2026!

Did you know that arXiv has an official merch shop? I didn’t, but I think it’s great. 😅 www.bonfire.com/store/arxiv-...

Dear @iclr-conf.bsky.social participants

It seems, that we have failed the communication about IMC26. Let's try again. The competition this year is here: kaggle.com/competitions... No prizes, but whole year leaderboard -- similar to KITTY and other academic competitions. 3D people, please retweet and share.

For people using superpoint for image matching, it might be interesting to know that I added a proof-of-concept steerer of superpoint here (making it possible to match rotated images with only one network pass per image): github.com/georg-bn/rot...

We're looking for a new colleague at @amlab.bsky.social: Assistant Professor in AI for Science 🔬🤖 World-class ML research, Amsterdam's thriving AI ecosystem (ELLIS, startups, big tech), and some of the best academic labor conditions in Europe ❤️ Deadline: May 30 👉 werkenbij.uva.nl/en/vacancies...

Sparse image matching is done via 1) keypoint detection in each image, 2) keypoint description, 3) matching of descriptions between images. Should rotation invariance be enforced at stage 2 or 3? Turns out both work fine! To be presented at the CVPR image matching workshop by @davnords.bsky.social

You don't imagine the future by mentally rendering a movie. You trace how things move -- abstractly, sparsely, step by step. We built a model that does exactly this. It predicts motion, not pixels -- and it's 3,000× faster than video world models. Myriad, accepted at @cvprconference.bsky.social

Happy Easter! Local Feature Matching has risen! arxiv.org/abs/2604.04931

My university (Chalmers University of Technology in 🇸🇪) is recruiting an assistant professor in data-driven cell & molecular biology, funded by the DDLS program @scilifelab.se #chemsky #facultychemjobs The position comes with a nice start-up package www.chalmers.se/en/about-cha...

Vibe graphing

Given the function specifying a sigmoid-MLP f(x)=Wₙσ(⋯σ(W₁x+b₁)⋯)+bₙ, we can recover the weights/biases up to (unique) signed permutations of the neurons by looking at f over the complex numbers, see the quoted thread. What happens if f is odd, i.e. f(x)=−f(−x)?

In 1994 Fefferman proved that by analytically continuing a neural network with sigmoid-activations to the complex numbers, you can read off the weights of the network by looking at the pole structure of the function. www.math.stonybrook.edu/~bishop/clas...

arxiv.org/abs/2603.23488 👀 I'll make a detailed thread later

🔮 Working on ML on curved manifolds? Don't miss out on Jacobi Fields! 🔮 I wrote a quick, highly visual and hopefully accessible introduction to the topic: "Jacobi Fields in Machine Learning" 🤠 Check it out here: olgatticus.github.io/blog/jacobi-...!

Today NeurIPS is announcing our official satellite event in Paris. After responding to the call from Ellis following the success of EurIPS in December, we are pleased to reach a new milestone by joining forces with the NeurIPS organizing committee for the 2026 edition.

I've made a SatAst, a small collection of hand-annotated satellite to astronaut image correspondences, public on github: github.com/georg-bn/sat.... This benchmark is part of the RoMa v2 paper, see Johan's thread below. bsky.app/profile/pars...

Enrollment form strikes again. (OpenReview -> Tasks -> ECCV -> Author Enrollment Form)

I hate the ECCV template so much. Why can't it just die? Nobody prints proceedings into LNCS volumes anymore. My favorite templates: 1. IEEE/CVPR 2. NeurIPS ... ... 39,2312,321. Edging tikz commands in morse code onto the skin of a dead walrus ... ... Springer/ECCV

I'm reviewing four papers. Three of them have a rebuttal. I wrote down my thoughts regarding the rebuttal for each. Across the three papers, one single (out of six) other reviewer has responded to the rebuttal.

What if position encodings were designed for vision from scratch? We introduce PaPE—Parabolic Position Encoding. Outperforms RoPE on 7/8 datasets and extrapolates to higher resolutions without fine-tuning or position interpolation. Paper, code, and website in thread 🧵

Our recent work looking into the desgining flexible approximately equivariant NNs. arxiv.org/abs/2602.02853 This is w/ Stefanos Pertigkiozoglou (stefanospert.github.io), Mircea Petrache, Kostas Daniilidis. Builds upon our work from the 2024 edition of NeurIPS proceedings.neurips.cc/paper_files/...

The #ECCV2026 Malmo 🇸🇪 call for papers is now available. Check it out! Call for Papers: eccv.ecva.net/Conferences/...

Looking forward to seeing the submission counts at ECCV and NeurIPS.

A chance to join us as postdoc in Gothenburg to workon this :) www.chalmers.se/en/about-cha...

Funny failure mode

Why you should probe more than just the final layer of your Vision Transformer to maximize performance. 🧵👇

Something @eugenevinitsky.bsky.social and I are very curious about... how can we make our client (a version of Bluesky for researchers) more friendly to grad students? What would encourage you all to post more?

Another Erdos problem this morning: (just to respond to a few people-- the system does NOT work by trying every possible answer and then checking. There's not enough matter and energy in the universe to solve theorems by trying every possible combination of symbols or whatever)

Power chords are even cooler in 5-part choir harmonization of a twelve tone row (from 2:17 open.spotify.com/track/7uezPJ... )

Reviewing for CVPR is sadly very boring.

New blog post (on a shiny new ICML blog!): What's New in #ICML2026 Peer Review Some highlights: - Policies to combat thinly sliced contributions - Cascading desk rejections for peer-review abuse - Reviewer reciprocity - New ways to support authors and reviewers Post: blog.icml.cc/2026/01/08/w...

I'd like to propose the following norm for peer review of papers. If a paper shows clear signs of LLM-generated errors that were not detected by the author, the paper should be immediately rejected. My reasoning: 1/ #ResearchIntegrity

every claim that "the incentives" support or deter certain kinds of behavior is also a statement about what kinds of external signals the claimant views as rewards or penalties #linklog

On the unexplained similarity across networks In behavior, order and weights, we keep seeing evidence that learning is more consistent than one might think. A walk through occurrences, my thoughts and the open question, why?! What's your hypothesis, missed papers and thoughts 🤖📈🧠 #AI

📢The second edition of ✨GRaM workshop✨ is here this time at #ICLR26. 🌟Submit your exciting works in Geometry-grounded representations. We welcome submissions in multiple tracks i.e. 📄 Proceedings 📝extended abstract 👩‍🏫Tutorial/blogpost as well as an exciting challenge!

Worried about AI’s military uses? We are too. We’re organising an ICLR 2026 workshop on AI research and military applications—dual-use risks, transparency, accountability, and ethical/legal governance & policy. Details + paper submissions: see Noa’s post and visit aiforpeaceworkshop.github.io.

An excellent article describing how LLMs fit into the bureaucratization of science: open.substack.com/pub/artifici...

You know what season it is! Right. Internship application season. Niantic Spatial is offering research internships on a multitude of 3D vision topics: relocalization, reconstruction, 3D VLMs... Top tier papers regularly come out of our internships 🚀 nianticspatial.careers.hibob.com/jobs/0fc4871...

#KostasKeynoteLessons: Curious about the "Keynote magic" behind my slides? I’m releasing the full Keynote source file for my recent Gaussian Splatting lecture, all 10 GIGAbytes of it! Grab the files in the thread and feel free to remix. Files: drive.google.com/drive/folder...

Super excited to bring GRaM 2.0 to ICLR 2026 in Brazil! Call for papers coming soon! @gram-org.bsky.social @iclr-conf.bsky.social

🧠🔬 Excited to share our #NeurIPS2025 paper: "Convolution Goes Higher-Order"! We asked: Can shallow networks be as expressive as deep ones? Inspired by biological vision, we introduce higher-order convolutions that capture complex image patterns standard CNNs miss. 🧵👇

🧠 How do neurons encode information? We know HOW MUCH, but what about WHAT information they encode? Our new work uses diffusion models to decompose neural information down to individual stimuli & features! 🎯Spotlight at #NeurIPS2025 🌟📄 arxiv.org/abs/2505.11309

I'm open to there being a role for blind review, but introducing non-blind review has a lot of upsides that may reduce how much we actually care about blind review. I think we care about blind review only because our publishing system is poorly designed and needs change in the modern era anyway.