Zhan Qu

About Me

Hi! I'm Zhan Qu, a Ph.D. student from AI4GC Lab at Zhejiang University, advised by Shengyu Zhang.

My current research focuses on GUI agent, especially on device-cloud collaboration: enabling lightweight on-device agents to perform robust GUI grounding and to autonomously decide when to request help from stronger cloud models at low cost. Our recent work, SPECTRA, adopts an active introspection paradigm by stress testing visual cognition and analyzing latent trajectories before decoding. Beyond my research work, I also lead the development of paper2anything, a toolkit of agent skills for academic paper promotion.

Previously, I worked on multi-agent collaboration and on the driving and rendering of 3D digital humans. Feel free to reach out for collaborations.

Projects

• paper2anything — An agent skills pack that turns an academic paper PDF into a presentation deck, conference poster, project homepage, Xiaohongshu post, or WeChat article. (code)