Yunhao (Andy) Ge (@GeYunhao) / X

Yunhao (Andy) Ge

141 posts

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

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
776
Yunhao (Andy) Ge
@GeYunhao
Dec 19, 2023
After completing an enriching and rewarding PhD journey, I am thrilled to announce that I am embarking on a new adventure as a Research Scientist at NVIDIA
20K
Yunhao (Andy) Ge
@GeYunhao
Dec 29, 2023
Inverse any images into prompt space for 2D/3D customized generation? DreamDistribution learns a prompt distribution of reference images, which can then be used to generate new 2D/3D instances capable of text-guided editing and more. Paper + Code + Web: briannlongzhao.github.io/DreamDistribut…
00:00
4.6K
Yunhao (Andy) Ge
@GeYunhao
Dec 11, 2023
3D Copy-Paste: seamlessly copy virtual objects and paste them into real scenes, maintaining physically plausible integration. This generated data enhances monocular 3D detection models, achieving State-of-the-Art performance. #NeurIPS2023 🌐 gyhandy.github.io/3D-Copy-Paste/
00:00
7.8K
Yunhao (Andy) Ge
@GeYunhao
Jul 30, 2023
How to allow AI That Teaches Other AI? [Our #TMLR paper] Shared Knowledge Lifelong Learning (SKILL) is a new lifelong learning paradigm that allows LL agents to share knowledge with each other; in the end, all agents master all tasks. 🌐 ilab.usc.edu/andy/skill
GIF
968
Yunhao (Andy) Ge
@GeYunhao
Jun 26, 2022
We have a new paper: DALL-E for Detection: Language-driven Context Image Synthesis for Object Detection arxiv.org/abs/2206.09592 A new paradigm for automatic context image generation at scale with DALL-E (generative models) for downstream takes (discriminative models).
Yunhao (Andy) Ge
@GeYunhao
Jun 21, 2023
How to make Vision Language Models reasonably admit "I do not know?" How to incorporate WordNet hierarchy into CLIP's decision-making process? Dive into our #CVPR poster at WED-AM-272: Improving Zero-Shot Generalization and Robustness of Multi-modal Models. 👇
928
Yunhao (Andy) Ge
@GeYunhao
Jun 2, 2023
Video for our CVPR 2023 paper: Improving Zero-shot Generalization and Robustness of Multi-modal Models @jessierenjie @balajiln @jiapingz Project Page: sites.google.com/usc.edu/hierar… Paper: arxiv.org/pdf/2212.01758… youtu.be/6nirYCh2xA0 via @YouTube
1.6K
Yunhao (Andy) Ge
@GeYunhao
Jul 5, 2022
#ECCV2022 Two papers on NeRF and Humanoid Neural Network are accepted by ECCV 2022, arXivs and code are coming soon. Thanks for all the co-authors!
Yunhao (Andy) Ge
@GeYunhao
Jun 21, 2023
Replying to @GeYunhao
[CVPR 2023 Hierarchy-CLIP] Improving Zero-Shot Generalization and Robustness of Multi-modal Models. project page: sites.google.com/usc.edu/hierar… paper: arxiv.org/pdf/2212.01758… code: github.com/gyhandy/Hierar… Youtube: youtube.com/watch?v=6nirYC… Bilibili: bilibili.com/video/BV1ih411…
194
Yunhao (Andy) Ge
@GeYunhao
Jul 30, 2023
How to allow AI That Teaches Other AI? Thanks! @USCViterbi viterbischool.usc.edu/news/2023/07/t… 🌐 ilab.usc.edu/andy/skill
AI That Teaches Other AI - USC Viterbi | School of Engineering
From viterbischool.usc.edu
324
Yunhao (Andy) Ge
@GeYunhao
Jan 7, 2024
Amazing work! Animal Kingdom series from @elliottszwu, Congrats @zizhang_li
Elliott / Shangzhe Wu
@elliottszwu
Jan 5, 2024
Nature presents a captivating confluence of similarity and diversity. Our new method 3D-Fauna learns a pan-category articulated 3D model of quadruped animals from Internet photos. At test time, it turns a single image into an animatable textured 3D mesh in a feed-forward pass.
00:00
640
Yunhao (Andy) Ge
@GeYunhao
Dec 12, 2023
Our #NeurIPS23 paper about 3D Copy-Paste was covered on USC news! Thanks @CSatUSC @USCViterbi! I will see you all at #NeurIPS23 Wednesday afternoon poster #305 and discuss more details of our 3D Copy-Paste: gyhandy.github.io/3D-Copy-Paste/
USC Thomas Lord Department of Computer Science
@CSatUSC
Dec 12, 2023
At #NeurIPS2023 this week! A new technique to “copy and paste” virtual 3D objects into real indoor scenes improving how computers see and interpret the world, led by @GeYunhao viterbischool.usc.edu/news/2023/12/c… @USCViterbi @33yuliangguo @jiajunwu_cs @StanfordAILab #AI
1.3K
Yunhao (Andy) Ge
@GeYunhao
Jun 19, 2021
CVPR 2021-A Peek Into the Reasoning of Neural Networks: Interpreting wit... youtu.be/ZzkpUrK-cRA via @YouTube This is the 5 minute oral video for our CVPR 2021 paper: Paper: arxiv.org/pdf/2105.00290… Code: github.com/gyhandy/Visual…...