Zesen (Hyman) Zhao @CVPR 2026
22 posts

Zesen (Hyman) Zhao @CVPR 2026
@SourORZ1
VLA Reserach Intern @GM. CSE PhD working on WAM and Robotics @Umich.


We (@saturdayrobotic) are hosting a Robotics & World Model happy hour at @CVPRConf 2026 on 6/6, fancy venue + delicious food. If you're into Robotics & World Models, join us: luma.com/zamm9g2g Lightning talks: 1. @neuralmotion — Introducing NM-GenET, a generative video-action model for universal embodiment transfer and cross-embodiment, cross-domain policy learning. 2. @nvidia Cosmos — Introducing Cosmos3, a next-generation omni world model that unifies image, video, audio, embodied reasoning, and robot policy control into a single scalable foundation model. 3. @XSquareRobot, @ZJU_China — WALL-WM, an event-centric World Action Model that scales general-purpose robot learning through semantic event pretraining, unified VLA inference, and large-scale real-world generalization. 4. @UMich — Exploring test-time scaling for World Action Models using zero-shot geometric verification to improve rollout quality, physical consistency, and downstream robot action performance without retraining. and more... See you in Denver @CVPR on 6/6!
























1/ 🎨✨ If an AI has never seen art before, can it develop artistic capabilities? Our team from MIT, Northeastern and ShanghaiTech University analyzes this question in our Art-Free Diffusion paper.






