Embodied AGI Group

team.png

Welcome to the UMass Embodied AGI Group!

Our goal is to develop intelligent agents capable of understanding and interacting with the world in a human-like manner. By combining physical and social intelligence with advanced models, we aim to push the boundaries of embodied general intelligence for real-world and virtual environments.

LGRC A101, 740 N. Pleasant Street, Amherst, MA 01003



Selected Publications

Physical Reasoning and Interaction

  1. RoboGen: Towards Unleashing Infinite Data for Automated Robot Learning via Generative Simulation
    Yufei Wang, Zhou Xian , Feng Chen , Tsun-Hsuan Wang , Yian Wang, Katerina Fragkiadaki, Zackory Erickson, David Held, and Chuang Gan
    arXiv PDF Code Website

Multimodal Language Models

  1. 3D-LLM: Injecting the 3D World into Large Language Models
    Yining Hong, Haoyu Zhen , Peihao Chen, Shuhong Zheng, Yilun Du , Zhenfang Chen, and Chuang Gan
    Advances in Neural Information Processing Systems
    arXiv PDF Code Website

Learning World Models

  1. 3D-VLA: A 3D Vision-Language-Action Generative World Model
    Haoyu Zhen, Xiaowen Qiu , Peihao Chen, Jincheng Yang, Xin Yan, Yilun Du, Yining Hong, and Chuang Gan
    arXiv PDF Code Website

Large Language Models

  1. Principle-Driven Self-Alignment of Language Models from Scratch with Minimal Human Supervision
    Zhiqing Sun, Yikang Shen , Qinhong Zhou, Hongxin Zhang , Zhenfang Chen, David Cox , Yiming Yang, and Chuang Gan
    Advances in Neural Information Processing Systems
    arXiv PDF Code Website

Audio-Visual Learning

  1. soundofpixels.png
    The Sound of Pixels
    Hang Zhao, Chuang Gan, Andrew Rouditchenko, Carl Vondrick, Josh McDermott, and Antonio Torralba
    In The European Conference on Computer Vision (ECCV)
    arXiv PDF Code Website