Xiongyi Cai

I'm a second-year MS student at UC San Diego, working with Prof. Xiaolong Wang. Before this, I received my BS in Electrical and Computer Engineering at SJTU, where I worked with Prof. Weidi Xie.

Email / X / Github / Google Scholar / LinkedIn

Research

I'm interested in computer vision, and robotics. My research aims to develop methods for learning skills from human data, enabling robots to complete dexterous tasks in real-world environments.

EgoVerse: An Egocentric Human Dataset for Robot Learning from Around the World
Ryan Punamiya*, Simar Kareer*, Zeyi Liu, Josh Citron, Ri-Zhao Qiu, Xiongyi Cai, Alexey Gavryushin, Jiaqi Chen, Davide Liconti, Lawrence Y. Zhu, Patcharapong Aphiwetsa, Baoyu Li, Aniketh Cheluva, Pranav Kuppili, Yangcen Liu, Dhruv Patel, Aidan Gao, Hye-Young Chung, Ryan Co, Renee Zbizika, Jeff Liu, Xiaomeng Xu, Haoyu Xiong, Geng Chen, Sebastiano Oliani, Chenyu Yang, Xi Wang, James Fort, Richard Newcombe, Josh Gao, Jason Chong, Garrett Matsuda, Aseem Doriwala, Marc Pollefeys, Robert Katzschmann, Xiaolong Wang, Shuran Song, Judy Hoffman, Danfei Xu
arXiv preprint, 2026
project page / arXiv / code /

This paper introduces EgoVerse, a large-scale egocentric human dataset and unified framework for scalable robot learning from egocentric human demonstrations.

In-N-On: Scaling Egocentric manipulation with in-the-wild and on-task data
Xiongyi Cai*, Ri-Zhao Qiu*, Geng Chen, Lai Wei, Isabella Liu, Tianshu Huang, Xuxin Cheng, Xiaolong Wang
arXiv preprint, 2025
project page / arXiv / code / model / data

This paper builds a large-scale egocentric human-humanoid manipulation dataset (PHSD) and uses it to pre-train and post-train a language-conditioned manipulation model, Human0.

Track-On: Transformer-based Online Point Tracking with Memory
Gorkay Aydemir, Xiongyi Cai, Weidi Xie, Fatma Guney
ICLR, 2025
project page / arXiv / code

This paper presents Track-On, a simple transformer-based model for online long-term point tracking.

Borrowed from Jon Barron's source code.