About Me
I am Zhiyu Zhou (ๅจๅฟๅฎ in Chinese), currently pursuing an M.Sc. in Artificial Intelligence at The Chinese University of Hong Kong (CUHK). I received my bachelorโs degree in Software Engineering from Jilin University, where I conducted research under the supervision of Prof. Hongxia Xie.
My research interests lie in robot learning, vision-language-action (VLA) models, and robotic manipulation, with a particular focus on the post-training of robot foundation models. I am especially interested in bridging perception, reasoning, and action to enable more generalizable and robust robotic behaviors.
My long-term goal is to develop general-purpose embodied intelligence that can reliably perceive, reason, and act in the real world.
Email: zhouzhiyu@link.cuhk.edu.hk
News
- 2026.08: Our work PinpointQA was accepted to ECCV EMR 2026.
- 2026.02: Our work MindPower was accepted to CVPR 2026.
Publications
PinpointQA: A Benchmark for Small Object-Centric Spatial Understanding in Indoor Videos
ECCV EMR 2026. PinpointQA evaluates small object-centric spatial understanding in indoor videos, spanning target presence, reference grounding, fine-grained description, and structured spatial prediction.
MindPower: Enabling Theory-of-Mind Reasoning in VLM-based Embodied Agents
CVPR 2026. MindPower explores Theory-of-Mind reasoning in vision-language-model-based embodied agents, aiming to improve how agents infer intentions, beliefs, and goals during embodied interaction.
Education
Expected
M.Sc. in Artificial Intelligence
B.Eng. in Software Engineering
Experience
PrimeBot
Affective Vision Computing (AVC) Lab, Jilin University
Honors and Awards
Jilin University