Incoming Ph.D. Student at the School of Computer Science, Peking University
Efficient inference and deployment systems for embodied intelligence.
I am an undergraduate in Computer Science and Technology at China University of Geosciences (Wuhan) and an incoming Ph.D. student at the School of Computer Science, Peking University, starting in Fall 2027. I currently work as a research intern in IF-Lab under the supervision of Prof. Xiang Chen.
I build efficient systems for Physical AI. My work focuses on making vision-language-action, vision-language-navigation, and world-action models fast, reliable, and deployable through speculative decoding, token caching, sparse computation, and edge-cloud collaboration.
My longer-term interest is in the systems behind long-horizon reasoning, physical consistency, and closed-loop robot control. I enjoy taking an idea from algorithm design to a running system and measuring whether it actually works on real hardware.
🤝 I am always happy to discuss research and open-source collaboration in embodied AI systems, efficient inference, and related infrastructure. Feel free to reach me at htxmzh@gmail.com.
Academic service: Reviewer for AAAI 2027, IJCNN 2027, and ICCAD 2027.
-
KERV: Kinematic-Rectified Speculative Decoding for Embodied VLA Models · DAC 2026
Kinematics-aware error compensation and adaptive acceptance for faster embodied VLA inference.
Paper · Code · -
APEX-SAM: Anatomy-aware Prompting with Expert Retrieval for Training-free Medical Image Segmentation · MICCAI 2026
Quality-aware expert retrieval, anatomy-aware prompting, and multi-modal fusion for cross-domain few-shot segmentation.
Project · Code · -
RAP: Retrieve, Adapt, and Prompt-Fit for Training-Free Few-Shot Medical Image Segmentation · IJCNN 2026
Retrieval and prompt fitting for training-free few-shot medical image segmentation.
Paper -
VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics Awareness · Preprint
Dynamics-aware token reuse for efficient vision-and-language navigation inference.
Paper
More publications and updates are available on my academic homepage and Google Scholar.
| Project | Role | Focus |
|---|---|---|
KERV  |
Core Contributor | Speculative decoding for embodied VLA models |
RoboECC  |
Core Contributor | Edge-cloud collaborative deployment for VLA models |
RoboNix  |
Core Contributor | Operating-system infrastructure and reusable services for embodied AI |
A-Pilot  |
Contributor | Mixed-precision quantization and edge deployment for MiniCPM-MoE |
CUG Template  |
Creator & Maintainer | PowerPoint, Beamer, Word, and LaTeX templates for CUG students |
I have also spent a great deal of time on competitive programming and robotics: ICPC Asia Regional bronze medals, a national second prize in RoboCup China, and a Meritorious Winner award in the Mathematical Contest in Modeling. These experiences shaped how I approach systems work: model the problem clearly, implement carefully, and make the result reproducible.