About Me
I am Tan Su (苏坦), an undergraduate student majoring in Information Engineering in the Department of Electrical and Electronic Engineering at the Southern University of Science and Technology (SUSTech). I am currently a Research Assistant at the Department of Computer Science and Engineering, The Chinese University of Hong Kong, under the supervision of Prof. James Cheng, where I study embodied AI and the generalization of generative robot policies. I am also an Algorithm Intern in the Foundation Model Group at KNOWIN AI, working closely with Dr. Binghui Xie on embodied spatial intelligence and multimodal robot-data pipelines. Previously, I was a Research Assistant with Prof. Ruogu Fang at the University of Florida and Prof. Lin Gu at The University of Tokyo, where I worked on generative medical imaging and controllable counterfactual fundus-image generation.
My research interests include embodied AI, diffusion-based generative models, and world models. My current work focuses on the generalization of generative robot policies and grounded spatial intelligence for embodied agents. I am actively seeking PhD or MPhil opportunities for Fall 2027 entry in Electrical Engineering, Computer Science, and Robotics. Please feel free to contact me by email or WeChat.
AI Researcher
- Embodied AI and robot learning
- Diffusion-based generative policies and world models
- Spatial perception, multimodal reasoning, and generalization
Football Player
- Captain of Zhicheng College Football Team
- Left winger, attacking midfielder, and defensive midfielder
- 32 appearances and 19 goals across college and university teams
Publications
Google Scholar
A General Bézier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis
A generative counterfactual framework that represents retinal vascular networks as controllable parametric structures.
arXiv
An Off-Line Self-Calibration Method for Resistive Insole Pressure Sensors
A lightweight neural calibration method for improving low-cost insole force measurements during real-world walking.
IEEE XploreEducation

Southern University of Science and Technology
B.Eng. candidate in Information Engineering
Department of Electrical and Electronic Engineering
Summer exchange at the University of Oxford, 2024.

The Affiliated High School of South China Normal University
University Preparatory Class
Research Experience
KNOWIN AI
Algorithm Intern · Foundation Model Group
Working closely with Dr. Binghui Xie on embodied spatial intelligence. I develop stereo-vision and multimodal data pipelines that convert real robot observations into grounded spatial-reasoning data for embodied vision-language models.
KNOWIN AI
The Chinese University of Hong Kong
Research Assistant · Department of Computer Science and Engineering
Working under the supervision of Prof. James Cheng on the generalization of generative robot policies. My current work develops controlled, closed-loop environments for studying how data coverage, action representations, and generative consistency affect policy performance.
Prof. James Cheng
University of Florida & The University of Tokyo
Research Assistant · First-Author Researcher
Worked with Prof. Ruogu Fang and Prof. Lin Gu on BTECF, a retinal-vessel counterfactual framework based on parametric Bézier trees. I independently built the complete codebase and end-to-end experimental pipeline, conducted all generative-model training, counterfactual and causal-analysis experiments, produced the figures, and led manuscript preparation for the first-author paper.
First-author paper
North Carolina State University
Summer Research Program Member
Worked with Prof. Zhishan Guo on cyber-physical system optimization for healthcare. I investigated diffusion-based lower-limb motion prediction from a minimal set of inertial and pressure-insole signals and compared it with LSTM and TCN baselines.
NC State University
Global Excellence Innovation Summer School
Team Lead · Autonomous Waste-Collection Vessel
Led a team in building a vision-guided surface vessel integrating waste detection, route planning, and robotic-arm collection. The project received the Silver Award.

Southern University of Science and Technology
Project Lead · SUSTech Innovation and Entrepreneurship Training Program
Leading a wearable muscle-monitoring project with Prof. Jiankun Wang and Prof. Qinghu Meng. I coordinate the team, optimize flexible temperature-sensor structures and placement, and integrate electromyography and humidity sensing to improve measurement robustness and wearability.

Southern University of Science and Technology
Project Lead · Co-First Author
Working with Prof. Chenglong Fu of the Department of Mechanical and Energy Engineering, I developed an off-line self-calibration method for low-cost resistive insole pressure sensors, built the wearable sensing system, and contributed as co-first author to the oral paper published at IEEE RCAR 2025.
IEEE RCAR paperFootball
Manchester City · France national team supporter
SUSTech University Football Team
Left winger · 2024 — Present
15 appearances · 10 goals
- 2024 Shenzhen City Tournament — Third Place · 3 goals in 3 matches
- 2025 Shenzhen City Tournament — Runner-up · 1 appearance
- 2025 Governor's Cup — Round of 16 · 1 goal in 4 matches
- 2026 Shenzhen Mengniu Cup — Runner-up · 2 goals in 4 matches
- 2026 PHBS Invitational — Champion · 4 goals in 3 matches

Zhicheng College Football Team
Captain · Left winger / Attacking midfielder / Defensive midfielder
17 appearances · 9 goals · 16 wins, 1 draw
- 2023 Freshman Cup — Champion · 2 goals in 4 matches
- 2024 College Cup — Champion · 5 goals in 4 matches
- 2025 College Cup — Champion · 3 appearances
- 2025 Freshman Cup — Runner-up · 1 goal in 2 matches
- 2026 College Cup — Champion · 1 goal in 4 matches

Awards
Second-Class Scholarship
Southern University of Science and Technology
Outstanding Student of the Academic Year
Southern University of Science and Technology
Silver Award · Shokz Global Excellence Innovation Summer School
Global Excellence Innovation Summer School
Guestbook
Messages are publicly attributed to the GitHub account used to post.
The board will begin scrolling after the first GitHub-attributed message is published.
Open the guestbook on GitHubGitHub attribution verifies the posting account, not a person’s legal identity. Select a message to enlarge it; select an author to visit their GitHub profile.
