Yunhao (Andy) Ge

@GeYunhao

Research Scientist @NVIDIA GEAR Lab | CS PhD @USC, Ex Visiting PhD @Stanford, Amazon ML Fellow @Amzaon, intern @Google, @Microsoft | VLA, World Foundation Model

Santa Clara

gyhandy.github.io

Joined January 2021

Pinned
Yunhao (Andy) Ge
@GeYunhao
Feb 4
Words in. Worlds imagined. Actions out. 🤖🌎 DreamZero lets robots dream in pixels and act—via joint video + action prediction. 🔥2× better generalization than VLAs ⚡14B @ 7 Hz 🤝Cross-embodiment transfer (w/ 10–20 min video) 🦾New robot, 30 min play, zero-shot skills intact
Joel Jang
@jang_yoel
Feb 4
Introducing DreamZero 🤖🌎 from @nvidia > A 14B “World Action Model” that achieves zero-shot generalization to unseen tasks & few-shot adaptation to new robots > The key? Jointly predicting video & actions in the same diffusion forward pass Project Page: dreamzero0.github.io
00:00
Yunhao (Andy) Ge
@GeYunhao
Jul 15
RoboTTT🤖 uses Test-Time Training with fast weights to tackle long-horizon robotic tasks, scaling visuomotor context to 8K timesteps—1,000× longer than prior policies—without increasing inference latency. Amazing work!🚀
Yunfan Jiang
@YunfanJiang
Jul 15
We scaled robot policies to 8K timesteps of visuomotor context, orders of magnitude beyond current SoTAs, at constant inference latency. Introducing RoboTTT 🤖 With minutes of experience in context, our robots: 🎥 one-shot imitate human video demos 📈 improve themselves during
00:00
Yunhao (Andy) Ge
@GeYunhao
Mar 23
Honored to contribute to DreamZero & Cosmos Policy 🤖 Seeing them featured in Jensen’s keynote made it all feel real — super proud of the team. GR00T N2, let’s GOOOO! 🚀
Yunhao (Andy) Ge
@GeYunhao
Feb 25
Human video is the most scalable source of physical intelligence. EgoScale answers a fundamental question: human video exhibits a clear scaling law for dexterous manipulation.
Jim Fan
@DrJimFan
Feb 25
We trained a humanoid with 22-DoF dexterous hands to assemble model cars, operate syringes, sort poker cards, fold/roll shirts, all learned primarily from 20,000+ hours of egocentric human video with no robot in the loop. Humans are the most scalable embodiment on the planet. We
00:00
Yunhao (Andy) Ge
@GeYunhao
Apr 24, 2025
DreamDistribution: Learning Prompt Distribution for Diverse In-distribution Generation 🎤 Catch @briannlongzhao at #ICLR2025 Poster Session 1 (#177) on Apr 24! 📄 Paper + Code + Demo: briannlongzhao.github.io/DreamDistribut…
Brian Zhao
@briannlongzhao
Apr 23, 2025
Introducing DreamDistribution, a novel, simple approach to learn semantic distribution over multiple input images, enabling generation of diverse in-distribution images and 3D with editing flexibilities. We will be presenting our work at @iclr_conf Apr 24 poster session 1 #177
00:00