#SIGGRAPHAsia2024 🔥 Replace your Marching Cubes code with our Occupancy-Based Dual Contouring to reveal the "real" shape from either a signed distance function or an occupancy function. No neural networks involved. Web: …pancy-based-dual-contouring.github.io

#NeurIPS2024 We'll be presenting SyncTweedies on Wednesday morning, a training-free diffusion synchronization technique that enables generation of various types of visual content using an image diffusion model. Wed, 11 a.m. - 2 p.m. PST East #2605 🌐 synctweedies.github.io

#NeurIPS2024 Thursday afternoon, don't miss Seungwoo Yoo's poster on Neural Pose Representation, a framework for pose generation and transfer based on neural keypoint representation and Jacobian field decoding. Thu 4:30 p.m. - 7:30 p.m. East #2202 🌐 neural-pose.github.io

#NeurIPS2024 DiT generates not only higher-quality images but also opens up new possibilities for improving training-free spatial grounding. Come visit Phillip (Yuseung) Lee 's GrounDiT poster to see how it works. Fri 4:30 p.m. - 7:30 p.m. East #2510 🌐 groundit-diffusion.github.io

#CVPR2025 🚀Check out **VideoHandles** by Juil (Juil Koo), the first method for test-time 3D object composition editing in videos. 🔗 Project: videohandles.github.io 📄 arXiv: arxiv.org/abs/2503.01107

🚀 Inference-time scaling for FLUX! Significant improvements in reward-guided generation with flow models, including text alignment, object counts, etc.—all at a compute cost under just $1! 📄 Paper: arxiv.org/abs/2503.19385 🔗 Project: flow-inference-time-scaling.github.io

Thanks AK! Our test-time technique makes image flow models way more controllable—better at matching text prompts, object counts, and object relationships; adding or removing concepts; and improving image aesthetics—all without finetuning! Project: flow-inference-time-scaling.github.io

ORIGEN Zero-Shot 3D Orientation Grounding in Text-to-Image Generation

Unconditional Priors Matter! The key to improving CFG-based "conditional" generation in diffusion models actually lies in the quality of their "unconditional" prior. Replace it with a better one to improve conditional generation! 🌐 unconditional-priors-matter.github.io

GPT-4o vs. Our test-time scaling with FLUX (1/2) GPT-4o still cannot count objects (see the ten, not seven, tomatoes on the left), but our test-time technique makes it work with FLUX. What you need is not a new model, but a test-time technique! 🌐 flow-inference-time-scaling.github.io

GPT-4o vs. Our test-time scaling with FLUX (2/2) GPT-4o cannot precisely understand the text (e.g., misinterpreting “occupying chairs” on the left), while our test-time technique generates an image perfectly aligned with the prompt. Check out more 👇 🌐 flow-inference-time-scaling.github.io

Introducing ORIGEN: the first orientation-grounding method for image generation with multiple open-vocabulary objects. It’s a novel zero-shot, reward-guided approach using Langevin dynamics, built on a one-step generative model like Flux-schnell. Project: origen2025.github.io

🚀 We’re hiring! The KAIST Visual AI Group is looking for Summer 2025 undergraduate interns. Interested in: 🌀 Diffusion / Flow / AR models (images, videos, text, more) 🧠 VLMs / LLMs / Foundation models 🧊 3D generation & neural rendering Apply now 👉 visualai.kaist.ac.kr/internship/

#ICLR2025 Come join our StochSync poster (#103) this morning! We introduce a method that combines the best parts of Score Distillation Sampling and Diffusion Synchronization to generate high-quality and consistent panoramas and mesh textures. stochsync.github.io

I recently presented our work, “Inference-Time Guided Generation with Diffusion and Flow Models,” at HKUST (CVM 2025 keynote) and NTU (MMLab), covering three classes of guidance methods for diffusion models and their extensions to flow models. Slides: onedrive.live.com/?redeem=aHR0cH…

Had an incredible opportunity to give two lectures on diffusion models at MLSS-Sénégal 🇸🇳 in early July! Slides are available here: onedrive.live.com/?redeem=aHR0cH… Big thanks to Eugene Ndiaye for the invitation!

Diffusion model course at SIGGRAPH 2025 is happening NOW in the West Building, Rooms 109–110. w/ Niloy Mitra, Or Patashnik, Daniel Cohen-Or, Paul Guerrero, and Juil Koo.

Great summary of the latest image & video diffusion models! Our #SIGGRAPH2025 course spans real-world uses to techniques like acceleration & flow matching. Slides: geometry.cs.ucl.ac.uk/courses/diffus… w/ Niloy Mitra, Or Patashnik, Daniel Cohen-Or , Paul Guerrero, Minhyuk Sung

Last month I presented our work on “Inference-Time Guided Generation w/ Diffusion & Flow Models” at NVIDIA, Google, Stanford, and SFU. I showed how recent flow matching models can be especially powerful for inference-time guidance. Check out the slides: drive.google.com/file/d/1zexSlw…

Five papers from our group have been accepted to #NeurIPS2025, including one spotlight! All were authored entirely by our members. The main focus is on test-time guided generation with diffusion and flow models, with one paper on neural PDE solving. More details to come.