Matheus Gadelha @gadelha.bsky.social · 10/10/2026Thanks for the translation -- this is a must read!! 010
Reposted by Matheus GadelhaGautam Kamath @gautamkamath.com · 18/08/2026Obviously this doesn't scale, but I chose to send this junior researcher a (hopefully polite) email with some of my thoughts (written hastily in 30 minutes, not polished). More practicing mathematicians & researchers should advocate for the culture they want, so here's my bit. 1234
Matheus Gadelha @gadelha.bsky.social · 05/08/2026I think I've never seen an instance of else being used instead of otherwise? Is this a tiktok thing? Actually the only way I clearly remember else being used in natural language is in a menacing tone; e.g. "You better do XYZ, or else!!" 110
Reposted by Matheus GadelhaChristian Wolf @chriswolfvision.bsky.social · 04/08/2026Rebuttals and discussions need to be reduced an extreme minimal length. CVPR style, 1 single PDF page. No discussion. Else, human ACs and reviewers will have a disadvantage compared to the world of agents out there now. 2233
Reposted by Matheus GadelhaShriram Krishnamurthi @shriram.bsky.social · 28/07/20260/ At the AIMCS (AI for Math and CS) workshop at FLoC 2026, @swarat.bsky.social ran a really fun panel on "How should the Formal Methods community respond to AI progress?" He put a lot of thought into very good questions, so here are my answers to them: ↵ 1145
Reposted by Matheus GadelhaYining Karl Li @yiningkarlli.bsky.social · 14/07/2026Woah, almost all of EGSR 2026 was recorded, and they posted it all to Youtube. Wow! Big props to the conference organizers for doing this! www.youtube.com/@egsr2026/pl...youtube.comEGSR 2026The 37th Eurographics Symposium on Rendering (EGSR 2026), Bordeaux (France). 0163
Reposted by Matheus GadelhaICML Conference @icmlconf.bsky.social · 10/07/2026We appreciate the love ❤️ But limited engagement and few followers on Bsky makes some (not us!) question whether it's worth it Be sure to follow (and share with your ML friends!) if you want to see more updates from ICML! 610924
Reposted by Matheus GadelhaGabriel Peyré @gabrielpeyre.bsky.social · 02/07/2026Slides for an introductory talk on IA for maths. speakerdeck.com/gpeyre/ia-fo...speakerdeck.comIA for theorySlide for an introduction talk on IA for maths. 1329
Matheus Gadelha @gadelha.bsky.social · 24/03/2026This work was led by our amazing intern Oindrila Saha (UMass Amherst -- now at Adobe!), with Vojtech Krs, Radomir Mech, Kevin Blackburn-Matzen, and Subhransu Maji (UMass Amherst). #ICLR2026 #ImageGeneration #ComputerVision 010
Matheus Gadelha @gadelha.bsky.social · 24/03/2026To train and evaluate at this scale, we introduce SIGMA-Set27K: a synthetic dataset with 100k+ unique subjects across 27k images, providing identity, structure, and spatial annotations. Training data at this level of annotation density did not previously exist for this task. 110
Matheus Gadelha @gadelha.bsky.social · 24/03/2026The framework also supports single- and multi-subject insertion in one pass, subject reposing, free-form masks not seen during training, and mixing different granularity levels in a single generation. 100
Matheus Gadelha @gadelha.bsky.social · 24/03/2026Subject identity is preserved while style varies freely via text. The same objects can be re-lit, re-stylized, or reposed without losing their appearance. 100
Matheus Gadelha @gadelha.bsky.social · 24/03/2026SIGMA-Gen also accepts spatial guidance at varying levels of precision — coarse 2D bounding boxes, 3D boxes, or pixel-level segmentation masks and depth maps — with a single model. You provide as much structure as you have; the model fills the rest. 100
Matheus Gadelha @gadelha.bsky.social · 24/03/2026The key challenge: placing multiple specific subjects in one scene while preserving each identity. Prior methods handle subjects independently or lose fidelity at compositing time. SIGMA-Gen does this in a single forward pass. Project page: oindrilasaha.github.io/SIGMA-Gen/arxiv.orgSIGMA-GEN: Structure and Identity Guided Multi-subject Assembly for Image GenerationWe present SIGMA-GEN, a unified framework for multi-identity preserving image generation. Unlike prior approaches, SIGMA-GEN is the first to enable single-pass multi-subject identity-preserved generat... 100
Matheus Gadelha @gadelha.bsky.social · 24/03/2026Text-to-image models are remarkable at generation, but they end up deciding on their own what goes where. If you need specific subjects in a specific arrangement, you are left writing prompts and sampling until something usable appears. SIGMA-Gen, our new ICLR 2026 paper, tries to change that. 160
Matheus Gadelha @gadelha.bsky.social · 15/03/2026I am very likely guilty of this myself, but I agree and really appreciate your take! 001
Matheus Gadelha @gadelha.bsky.social · 04/03/2026PSA (this sounds like something pretty important that I should know but could have lived the rest of my life in shame without ever learning it) 010
Reposted by Matheus GadelhaIgor Gilitschenski @igilitschenski.bsky.social · 13/02/2026🚀 Excited to share REPPO, a new on-policy RL agent! TL;DR: Replace PPO with REPPO for fewer hyperparameter headaches and more robust training. REPPO, led by @cvoelcker.bsky.social, will be presented at ICLR 2026. How does it work? 🧵👇 12510
Matheus Gadelha @gadelha.bsky.social · 09/01/2026Great resource! This is specially important for people starting their research careers. Getting to know different research communities is really important! Technical insight is crucial, but science is ultimately a social endeavor -- understanding and communicating to your peers is key. 131
Reposted by Matheus GadelhaAaron Hertzmann @aaronhertzmann.com · 26/11/2025As we in the US gather with our families this weekend, it's a good time to consider the ways that technology brings us together—but also keeps us apart. In this blog post, I tell a story about how technology may have tended to increase social isolation. aaronhertzmann.com/2025/10/26/i...aaronhertzmann.comTechnology and Social Isolation: From Cars to “AI”In this blog post, I tell a story of how some technologies from the past century or so have, overall, led to increased social isolation in the United States, as we replaced in-person social interactio... 074
Reposted by Matheus GadelhaAlexander Doria @dorialexander.bsky.social · 10/11/2025Breaking: we release a fully synthetic generalist dataset for pretraining, SYNTH and two new SOTA reasoning models exclusively trained on it. Despite having seen only 200 billion tokens, Baguettotron is currently best-in-class in its size range. pleias.fr/blog/blogsyn... 318833
Matheus Gadelha @gadelha.bsky.social · 23/10/2025Session this afternoon (in 30 minutes)!! Poster 153 — see you there! 020
Reposted by Matheus GadelhaDavid Picard @davidpicard.eurosky.social · 27/06/2025I wrote a notebook for a lecture/exercice on image generation with flow matching. The idea is to use FM to render images composed of simple shapes using their attributes (type, size, color, etc). Not super useful but fun and easy to train! colab.research.google.com/drive/16GJyb... Comments welcome! 2418
Matheus Gadelha @gadelha.bsky.social · 20/06/2025Are the results out? I see nothing in OpenReview :-( 210
Matheus Gadelha @gadelha.bsky.social · 11/06/2025For folks attending CVPR: is there a website where I can see the list of workshops, their location AND time? Day and time are empty when I access cvpr.thecvf.com/Conferences/...cvpr.thecvf.comCVPR 2025 Workshop List 010
Matheus Gadelha @gadelha.bsky.social · 09/06/2025I will be in Nashville until Saturday for CVPR'25 \o/ DM if you want to meet! 060
Reposted by Matheus GadelhaÉlie Michel @elie-michel.bsky.social · 19/05/2025🏅Honored to have been awarded at #Eurographics25 for our paper on #LipschitzPruning to speed-up SDF rendering! 👉 The paper's page: wbrbr.org/publications... Congrats to @wbrbr.bsky.social, M. Sanchez, @axelparis.bsky.social, T. Lambert, @tamyboubekeur.bsky.social, M. Paulin and T. Thonat! 0254
Matheus Gadelha @gadelha.bsky.social · 14/05/2025I usually write a python script that prints some .npy file as tex a .tex table. It is also useful as an easy way to share results throughout the project, so I consider this as part of the codebase. I heard that people that are more serious about such practice use smth like github.com/JelteF/PyLaTeXgithub.comGitHub - JelteF/PyLaTeX: A Python library for creating LaTeX filesA Python library for creating LaTeX files. Contribute to JelteF/PyLaTeX development by creating an account on GitHub. 020
Matheus Gadelha @gadelha.bsky.social · 14/05/2025Comecei a fazer por causa disso e por causa do vim, mas até no overleaf é vantagem 000
Matheus Gadelha @gadelha.bsky.social · 14/05/2025Eu n uso pra coisas que faço sozinho (anotações, apresentações, etc). Mas pra trabalho colaborativo é meio q obrigatório. Estudantes vão entrar em revolta se vc usar git hahaha 100
Matheus Gadelha @gadelha.bsky.social · 14/05/2025NeurIPS and SIGGRAPH Asia deadline are coming. Make your life easier: read this thread. 172
Matheus Gadelha @gadelha.bsky.social · 28/04/2025Let's gooo!!! \o/ Probably my first time visiting Brazil for professional reasons :-) 040
Matheus Gadelha @gadelha.bsky.social · 15/04/2025What features did you find particularly useful? I liked asking questions about the code base and the tab completion seems nice, but I've been getting unhelpful suggestions for all the "agentic" stuff. 100
Reposted by Matheus Gadelha#CVPR2026 @cvprconference.bsky.social · 27/02/2025By popular demand, we are extending #CVPR2025 coverage to Bluesky. Stay tuned! 512417
Reposted by Matheus GadelhaAngjoo Kanazawa @akanazawa.bsky.social · 24/02/2025Exciting news! MegaSAM code is out🔥 & the updated Shape of Motion results with MegaSAM are really impressive! A year ago I didn't think we could make any progress on these videos: shape-of-motion.github.io/results.html Huge congrats to everyone involved and the community 🎉 37417
Matheus Gadelha @gadelha.bsky.social · 23/02/2025I understand the sentiment, but it is important for people to know that is currently does not reflect reviewer guidelines at CVPR: cvpr.thecvf.com/Conferences/... “(…) you should include specific feedback on ways the authors can improve their papers.”cvpr.thecvf.com2025 Reviewer Guidelines 250
Matheus Gadelha @gadelha.bsky.social · 21/02/2025But is Meta indeed cancelling professional fact checking worldwide? Their response to the Brazilian supreme court inquiry said the dismissal of professional fact checking was exclusive to the US 000
Reposted by Matheus GadelhaQianqian Wang @qianqianwang.bsky.social · 18/02/2025Late to post, but excited to introduce CUT3R! An online 3D reasoning framework for many 3D tasks directly from just RGB. For static or dynamic scenes. Video or image collections, all in one! Project Page: cut3r.github.io Code and Model: github.com/CUT3R/CUT3R 2346
Matheus Gadelha @gadelha.bsky.social · 13/02/2025It’s a plugin for neovim that has some stuff similar to cursor; it is open source and you can configure it to use the model of your choice 010
Matheus Gadelha @gadelha.bsky.social · 13/02/2025I am using Avante + vim and I am pretty happy about it 120
Reposted by Matheus GadelhaHarry Thasarathan @hthasarathan.bsky.social · 07/02/2025🌌🛰️🔭Wanna know which features are universal vs unique in your models and how to find them? Excited to share our preprint: "Universal Sparse Autoencoders: Interpretable Cross-Model Concept Alignment"! arxiv.org/abs/2502.03714 (1/9) 15617
Reposted by Matheus GadelhaAaron Hertzmann @aaronhertzmann.com · 06/02/2025New paper about pictures: I identify trends in geometric perspective in my own drawings and photos, and compare them to how the original scenes looked. I discuss what these trends might say about art history and vision science. Published in _Art & Perception_. #visionscience psyarxiv.com/pq8nbpsyarxiv.comOSF 3196
Reposted by Matheus GadelhaAndrea Tagliasacchi @taiyasaki.bsky.social · 05/02/2025"𝐑𝐚𝐝𝐢𝐚𝐧𝐭 𝐅𝐨𝐚𝐦: Real-Time Differentiable Ray Tracing" A mesh-based 3D represention for training radiance fields from collections of images. radfoam.github.io arxiv.org/abs/2502.01157 Project co-lead by my PhD students Shrisudhan Govindarajan and Daniel Rebain, and w/ co-advisor Kwang Moo Yi 25412
Matheus Gadelha @gadelha.bsky.social · 03/02/2025It would be nice to mention someone and have them receive an e-mail with a subject like "You (Reviewer HjKl) were mentioned in a discussion for Paper 447809 in OpenReview" 000
Matheus Gadelha @gadelha.bsky.social · 03/02/2025#CVPR2025 Is there any way in OpenReview to "mention" a reviewer in a discussion? I think reviewers get the e-mail with whatever message gets posted in the discussion that is sent to them, but they have no idea if Reviewer HjKl is them or someone else... 110
