Boyao Han

Boyao Han

I am an undergraduate student (class of 2022) at the College of Computer Science and Electronic Engineering, Hunan University. I am currently a research assistant at the School of Data Science, The Chinese University of Hong Kong, Shenzhen, advised by Prof. Li Jiang. My research interests lie in Embodied AI and Computer Vision, with a focus on Vision-Language-Action / Video-Action Models, 3D Computer Vision, and multimodal learning.

Publications

You can also find my articles on my Google Scholar profile.

Conference Papers

GeoPredict: Leveraging Predictive Kinematics and 3D Gaussian Geometry for Precise VLA Manipulation

Published in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Highlight, 2026

GeoPredict is a geometry-aware VLA framework that injects predictive kinematic trajectories and 3D Gaussian geometry priors as training-time supervision, enabling precise 3D reasoning for robotic manipulation without extra inference cost. It consistently outperforms strong VLA baselines on RoboCasa Human-50, LIBERO, and real-world geometry-intensive tasks.

Recommended citation: Jingjing Qian, Boyao Han, Chen Shi, Lei Xiao, Long Yang, Shaoshuai Shi, Li Jiang. (2026). "GeoPredict: Leveraging Predictive Kinematics and 3D Gaussian Geometry for Precise VLA Manipulation." CVPR. (Highlight).
Download Paper

PointSLAM++: Robust Dense Neural Gaussian Point Cloud-based SLAM

Published in AAAI Conference on Artificial Intelligence (AAAI), Poster, 2025

PointSLAM++ is an RGB-D SLAM system that combines a hierarchically constrained neural Gaussian representation, progressive pose optimization, and a dynamic neural representation graph to achieve robust, high-precision real-time 3D reconstruction and photorealistic rendering under depth-sensor noise.

Recommended citation: Xu Wang, Boyao Han, Xiaojun Chen, Ying Liu, Ruihui Li. (2026). "PointSLAM++: Robust Dense Neural Gaussian Point Cloud-based SLAM." AAAI. (Poster).
Download Paper

Memory Forcing: Spatio-temporal Memory for Consistent Scene Generation on Minecraft

Published in arXiv preprint, 2025

Memory Forcing is a training framework that enables video diffusion models to balance exploration and revisit consistency for autoregressive video generation in Minecraft using spatio-temporal memory mechanisms.

Recommended citation: Junchao Huang, Xinting Hu, Boyao Han, Shaoshuai Shi, Zhuotao Tian, Tianyu He, Li Jiang. (2025). "Memory Forcing: Spatio-temporal Memory for Consistent Scene Generation on Minecraft." arXiv:2510.03198.
Download Paper