Very happy to share my first project at NVIDIA! 🚀 We present Meshtron, a pure Transformer model that consistently reasons over 500k+ tokens to generate 3D meshes of outstanding quality. Webpage: research.nvidia.com/labs/dir/mesht… w/ @zekunhao19951 Tsung-Yi Lin & Ming-Yu Liu 🧵

🔥 3/4 #ICLR2025 Grounding Continuous Representations in Geometry: Equivariant Neural Fields (ENF), for geometry-informed continuous signal representations. Learn more: openreview.net/forum?id=A4eCz… with D. Wessels, David M. Knigge Samuele Papa Riccardo Valperga Sharvaree Vadgama Erik Bekkers

Our team is actively recruiting at all levels. Come join us in this exciting journey! 🚀

✨ The VIS Lab at the @University of Amsterdam is proud and excited to announce it has #TWELVE papers  🚀  accepted for the leading #AI-#makers conference on representation learning (#ICLR) in Singapore 🇸🇬. 1/n 👇👇👇

AI/ML Internships in Drug Discovery🚨 Our team is hiring PhD research interns for summer 2025 in the US. Come work with us on cutting-edge drug discovery projects! We have 3 openings: 1. Hyperbolic NNs for sequence modeling: jobs.jnj.com/en/jobs/250623… 2. Causal Inference and

🚀New paper at #ICLR2025!🚀 We introduce Equivariant Neural Fields (ENFs): a new class of Conditional Neural Fields (CNFs) that explicitly ground continuous representations in geometry across image & 3D shapes reconstrution, PDE forecasting & more 🧵Read on for details! 👇

Variational Flow Matching goes Riemannian! 🔮 In this preliminary work, we derive a variational objective for probability flows 🌀 on manifolds with closed-form geodesics. My dream team: Floor Eijkelboom Alison Erik Bekkers 💥 📜 arxiv.org/abs/2502.12981 🧵1/5

Join Us for an Exciting Hackathon on Equivariant Neural Fields (ENFs) & Geometry-Grounded Representations! ✨ The Deep Thinking Hour, ELLIS unit Amsterdam & the New Theory joint forces, and host a hackathon focused on ENFs and geometry-grounded representations

We are looking for an intern to join MLR at Apple ASAP. You will be working with me, Yuyang Wang and other scientists in the team on domain-agnostic generative modeling at scale. If you are interested please reach out to mbautistamartin at apple by the end of this week.

We’re looking for an intern at Apple MLR 🍎 ASAP. Join us if interested in building universal diffusion/flow-matching model at scale!

🤹 Excited to share Erwin: A Tree-based Hierarchical Transformer for Large-scale Physical Systems joint work with Max Welling and Jan-Willem van de Meent Core components of Erwin: - hierarchical organization of data via ball trees - localized attention for linear complexity - hardware-efficient

🔥 CuspAI is on fire! We're on a mission to revolutionize the world with AI-designed materials that can change the game. 🤖 We’re thrilled to welcome our new board members: Yann LeCun , Kristin Persson and Verity Harding 🚀 Check out the news: fortune.com/2025/03/05/cus… 1/

Super happy to share our work was accepted as an Oral at the Delta Workshop @ ICLR 2025! 🎉 Can’t wait to talk about it in Singapore 😎 Congrats to the amazing team Floor Eijkelboom Alison Erik Bekkers 🔥

We are happy to introduce Cosmos-Reason1, our reasoning model specifically trained for Physical AI. Really proud of being part of this immense effort 💪

So excited for New Theory's first open hire! Their approach—building world models via symmetry/geometry, rooted in the foundational principles of nature & intelligence—is, I believe, the *only* way forward for AI to truly grasp our world. Foundational work! 👇

✨ Job opportunity!! ✨ Interested in working for a startup building the next gen of scalable AI using Geometric Deep Learning? Check out New Theory's opening for Sr AI Researcher! Interning here currently, loving every second of it. The team and their ideas are awesome🥰🚀

🚀🚀🚀Our paper "Probabilistic Interactive 3D Segmentation with Hierarchical Neural Processes" has been accepted to ICML 2025! A small but exciting step toward interpretable and interpretable Human-AI Interaction. Find more details in our project page: jliu4ai.github.io/NPISeg3D_proje…

🌍 From earthquake prediction to robot navigation - what connects them? Eikonal equations! We developed E-NES: a neural network that leverages geometric symmetries to solve entire families of velocity fields through group transformations. Grid-free and scalable! 🧵👇

We've raised $100M from Kleiner Perkins, Index Ventures, Lightspeed, and NVIDIA. Today we're introducing Sonic-3 - the state-of-the-art model for realtime conversation. What makes Sonic-3 great: - Breakthrough naturalness - laughter and full emotional range - Lightning fast -

Cool news: our extended Riemannian Gaussian VFM paper is out! 🔮 We define and study a variational objective for probability flows 🌀 on manifolds with closed-form geodesics. Floor Eijkelboom Alison Cong Liu Max Welling Jan-Willem van de Meent Erik Bekkers 🔥 📜 arxiv.org/abs/2502.12981