*動画生成：Sunoで作ったオリジナル曲をComfyUI-FLOATで歌ってもらう #comfyui note投稿しました。 note.com/bakushu/n/n1f8…

DTM vs FM👇 Lots of interest in how Difference Transition Matching (DTM) connects to Flow Matching (FM). Here is a short animation that illustrates Theorem 1 in our paper: For a very small step size (1/T), DTM converges to an Euler step of FM.

Thank you for sharing our work!

Some news: We're building the next big thing — the first-ever AI-only social video app, built on a highly expressive human video model. Over the past few weeks, we’ve been testing it in private beta. Now, we’re opening early access: download the iOS app to join the waitlist, or

Our Agent Distillation paper is accepted at #NeurIPS2025 Spotlight! 🚀 Turn your small LM into a strong agent 💪 Code: github.com/Nardien/agent-…

I'll be at NeurIPS to present the final paper of my PhD: Continuous Diffusion Model for Language Modeling (arxiv.org/abs/2502.11564) We present a continuous diffusion model for language modeling using tools from Riemannian geometry, opening up a new direction for diffusion LMs!

I'm excited to be presenting our work at NeurIPS 2025! 🗓 When: Wed, Dec 3, 2025 11:00 AM - 2:00 PM (PST) 📍Where: Exhibit Hall C, D, E #1606 If you're attending NeurIPS, please stop by. I'd love to chat about the work and AI4Science!

Avatar Forcing A real-time interactive head avatar generation framework that enables natural conversation with 500ms latency. Uses diffusion forcing for causal motion generation and direct preference optimization for expressive interactions without labeled data.

We just upgraded XLeRobot 🚀 Built by the MakerMods team Isaac Sin, Mr Thompson and QI LIU. • Easier to build • Improved chassis • Reduced 3D-print time and material • Designed in collaboration with the original author Vector Wang Fully open-source Full build

Self-Refining Video Sampling: inference-time method using a video generator as its own refiner to correct physics and motion. no retraining needed; scores >70% human preference; is validated on Wan2.2 & Cosmos. agwmon.github.io/self-refine-vi…

What if your video generator could refine itself—at inference time? ❌No new models. ❌No retraining. ❌No external verifier. 💡 Introducing Self-Refining Video Sampling By reinterpreting a pretrained generator (Wan2.2, Cosmos) as a denoising autoencoder, we enable iterative

if you are building video diffusion / world simulators, try this new sampler. temporal consistency pins videos to a low-dimensional manifold in the total pixel space. self-refinement sampling keeps them there.

Learning from robot data? Standard. Direct Video-Action Models (DVA) is different: treat robot control as video generation, then translate the generated video into actions. Built by , the system pre-trains causal video models from scratch and can run complex

🎾Introducing LATENT: Learning Athletic Humanoid Tennis Skills from Imperfect Human Motion Data Dynamic movements, agile whole-body coordination, and rapid reactions. A step toward athletic humanoid sports skills. Project: zzk273.github.io/LATENT/ Code: github.com/GalaxyGeneralR…

We developed an RL method for fine-tuning our models for precise tasks in just a few hours or even minutes. Instead of training the whole model, we add an “RL token” output to π-0.6, our latest model, which is used by a tiny actor and critic to learn quickly with RL.

"Diffusability" is all about the spectrum. arxiv.org/abs/2603.14645 If you enjoyed my blog post about diffusion as spectral autoregression, and are wondering how this relates to latent diffusion, give this paper a read!