Reconstructing dynamic scenes from video inputs is challenging due to the sparse nature of inputs, both in time and space. In our new preprint, we address this issue by introducing the ReMatching framework - a novel approach for designing and integrating deformation priors into

Our **Flow Matching Tutorial** from #NeurIPS2024 is now publicly available: neurips.cc/virtual/2024/t… Heli Ben-Hamu Ricky T. Q. Chen

📣I'll be at the poster session with our follow-up on Discrete Flow Matching. We derive a closed-form solution to the kinetic optimal problem for conditional velocity on discrete spaces. Into flow models? come chat! 💬 🗓Poster: Sat 10am (#191), 🎤Oral: Sat 3:30pm (6E) #ICLR2025

What makes a good 3D scene representation? Instead of meshes or Gaussians, we propose Superquadrics to decompose 3D scenes into extremely compact representations ➡️ check out our paper for exciting use-cases in robotics🤖 and GenAI🚀 super-dec.github.io w/ Elisabetta Fedele Marc Pollefeys

Excited to share our recent work on corrector sampling in language models! A new sampling method that mitigates error accumulation by iteratively revisiting tokens in a window of previously generated text. With: Neta Shaul Uriel Singer Yaron Lipman Link: arxiv.org/abs/2506.06215

Had a great chat with Neta Shaul a few days ago about transition matching. Apparently, breaking the generation process into multiple intermediate generation processes really boosts both quality and performance — check it out 👇🏻👇🏻

Excited to share Flow Matching Policy Gradients: expressive RL policies trained from rewards using flow matching. It’s an easy, drop-in replacement for Gaussian PPO on control tasks.

We’ve been building something we’re 𝑟𝑒𝑎𝑙𝑙𝑦 excited about – LL3M: LLM-powered agents that turn text into editable 3D assets. LL3M models shapes as interpretable Blender code, making geometry, appearance, and style easy to modify. 🔗 threedle.github.io/ll3m 1/

Introducing 𝗦𝗧𝗿𝗲𝗮𝗺𝟯𝗥, a new 3D geometric foundation model for efficient 3D reconstruction from streaming input. Similar to LLMs, STream3R uses casual attention during training and KVCache at inference. No need to worry about post-alignment or reconstructing from scratch.

My acceptance speech at the Turing award ceremony: Good evening ladies and gentlemen. The main idea of reinforcement learning is that a machine might discover what to do on its own, without being told, from its own experience, by trial and error. As far as I know, the first

Excited to share our work Set Block Decoding! A new paradigm combining next-token-prediction and masked (or discrete diffusion) models, allowing parallel decoding without any architectural changes and with exact KV cache. Arguably one of the simplest ways to accelerate LLMs!

Can we use video diffusion to generate 3D scenes? 𝐖𝐨𝐫𝐥𝐝𝐄𝐱𝐩𝐥𝐨𝐫𝐞𝐫 (#SIGGRAPHAsia25) creates fully-navigable scenes via autoregressive video generation. Text input -> 3DGS scene output & interactive rendering! 🌍mschneider456.github.io/world-explorer/ 📽️youtu.be/N6NJsNyiv6I

How do we reconstruct a 3D scene from photos with varying exposures? Standard methods often fail, leaving you with blown-out colors or disturbing shadows. We're excited to introduce Neural Exposure Fields (NExF), our new work accepted at #NeurIPS2025! 🧵

truly the greatest day ever🎗️

Michal Irani receiving the distinguished researcher award.

Had a blast talking about Transition Matching at the HUJI Vision Seminar, big thanks to Eliahu Horwitz for inviting me! 🚀 If you like simple visual illustrations of complex ideas, I made a few in my slides: neta93.github.io/slides/transit…

How can we standardize conditioning signals in image/video models to achieve the iterative editing & portability that Universal Scene Descriptors provide in graphics? Introducing Neural USD: An object-centric framework for iterative editing & control 🧵

Introducing XFactor: the first pose- and geometry-free method capable of true Novel View Synthesis (NVS). We re-think NVS and the concept of camera poses completely without concepts from multi-view geometry as a pure representation learning problem! mitchel.computer/xfactor/ (1/n)

We present Time-to-Move (TTM)! a training-free, plug-and-play method for precise motion control in video diffusion. Unlike prior training-based methods, TTM works with any backbone at no extra cost🔥 Page: time-to-move.github.io [1/4] Noam Rotstein Or Litany Amir Mann