AffordGen: Generating Diverse Demonstrations for Generalizable Object Manipulation with Affordance Correspondence

Jiawei Zhang*, Kaizhe Hu*, Yingqian Huang, Yuanchen Ju, Zhengrong Xue, Huazhe Xu (* equal contribution)

CVPR 2026 2026

AffordGen scales a small number of source demonstrations into affordance-grounded manipulation trajectories across novel objects, categories, and full 6D pose variations, enabling closed-loop visuomotor policies to generalize to unseen objects in simulation and the real world.

[Project] [Paper] [arXiv] [Code]

OS-MAP: How Far Can Computer-Using Agents Go in Breadth and Depth?

Xuetian Chen, Yinghao Chen, Xinfeng Yuan, Zhuo Peng, Lu Chen, Yuekeng Li, Zhoujia Zhang, Yingqian Huang, Leyan Huang, Jiaqing Liang, Tianbao Xie, Zhiyong Wu, Qiushi Sun, Biqing Qi, Bowen Zhou

ICML 2025 Workshop on Computer Use Agents (WCUA), Poster 2025

OS-MAP is a benchmark for computer-using agents that organizes 416 realistic tasks across 15 desktop applications by automation level and generalization scope, enabling structured analysis of agent capabilities.

[Paper] [arXiv] [Code] [OpenReview]