Ph.D. in Computer Science
Texas A&M University
Xiangbo Gao
Generative AI researcher working on video generation, controllable video editing, and world models.
Profile
About me
I am a Computer Science Ph.D. student at Texas A&M University. My research focuses on generative video models, controllable video editing, and world models, with an emphasis on precise and physically grounded visual generation.
- LocationTexas A&M University, College Station, TX
- ResearchVideo generation, controllable editing, and world models
- InterestsSnowboarding, Skiing, Rock climbing, Badminton, Tennis
- AffiliationDepartment of Computer Science & Engineering, Texas A&M University
Academic path
Education
2025.1 - Present
M.S. in Robotics
University of Michigan, Ann Arbor
B.S. in Computer Science | B.S. in Mathematics
University of California, Irvine
Industry experience
Employment
Research Intern
Adobe Research
Autonomous Driving Algorithms Research Intern
DiDi Global Inc., USA
Perception Research Intern
Anhui Cowa ROBOT Co., Ltd, Shanghai, China
Research
Publications
Selected work in generative video, controllable editing, world models, and autonomous systems.
Selected
7 works
ECCV 2026
SparkVSR: Interactive Video Super-Resolution via Sparse Keyframe Propagation
Jiongze Yu, Xiangbo Gao, Pooja Verlani, Akshay Gadde, Yilin Wang, Balu Adsumilli, Zhengzhong Tu
An interactive video super-resolution framework that propagates sparse user-selected high-resolution keyframes across a video.
CVPR 2025 Workshop ยท Best Paper Award
LangCoop: Collaborative Driving with Language
Xiangbo Gao, Yuheng Wu, Runsheng Wang, Chenxi Liu, Yang Zhou, Zhengzhong Tu
A language-driven vehicle collaboration framework that cuts communication bandwidth by 96% while preserving closed-loop driving performance.
TMLR 2026
AutoTrust: Benchmarking Trustworthiness in Large Vision Language Models for Autonomous Driving
Shuo Xing, Hongyuan Hua, Xiangbo Gao, Shenzhe Zhu, Renjie Li, Kexin Tian, Xiaopeng Li, Heng Huang, Tianbao Yang, Zhangyang Wang, Yang Zhou, Huaxiu Yao, Zhengzhong Tu
A comprehensive benchmark for evaluating the trustworthiness of large vision-language models in autonomous driving.
ICASSP 2024
Scale-free and Task-agnostic Attack: Generating Photo-realistic Adversarial Patterns with Patch Quilting Generator
Xiangbo Gao, Cheng Luo, Qinliang Lin, Weicheng Xie, Minmin Liu, Linlin Shen, Keerthy Kusumam, Siyang Song
A scale-free generator for transferable, defense-resistant, and visually realistic adversarial patterns.
MICCAI 2022
Sample Hardness Based Gradient Loss for Long-Tailed Cervical Cell Detection
Minmin Liu, Xuechen Li, Xiangbo Gao, Junliang Chen, Linlin Shen, Huisi Wu
Gradient-based hardness calibration improves long-tailed cervical cell detection by 7.8% mAP over standard classification loss.
Preprints
8 works
arXiv 2026
Visko Orbis 1.0: A Live Model for Real-Time Interactive Long Video Generation
Xiangbo Gao, Siyuan Yang, Ping He, Mingyang Wu, Yuheng Wu, Yushen Zuo, Jiongze Yu, Ryan Cui, Hongyuan Hua, Devin Ma, Xiao Jin, Yubo Yuan, Qing Yin, Jie Yang, Zhengzhong Tu
A live video model for prompt-switchable, hour-scale generation with real-time 4K output at 24 FPS.
arXiv 2026
Neuro-Symbolic Drive: Rule-Grounded Faithful Reasoning for Driving VLAs
Xiangbo Gao, Xiukun Huang, Boyu Lu, Junge Zhang, Mengjie Mao, Jiachen Li, Wei Xiong, Zhengzhong Tu
Rule-based planners provide faithful reasoning traces that keep a driving VLA's explanations structurally tied to its motion decisions.
arXiv 2026
Delta Forcing: Trust Region Steering for Interactive Autoregressive Video Generation
Yuheng Wu, Xiangbo Gao, Tianhao Chen, Xinghao Chen, Qing Yin, Zhengzhong Tu, Dongman Lee
A trust-region training objective improves long-horizon consistency while keeping interactive video generation responsive to new events.
arXiv 2026
VEFX-Bench: A Holistic Benchmark for Generic Video Editing and Visual Effects
Xiangbo Gao, Sicong Jiang, Bangya Liu, Xinghao Chen, Minglai Yang, Siyuan Yang, Mingyang Wu, Jiongze Yu, Qi Zheng, Haozhi Wang, Jiayi Zhang, Jared Yang, Jie Yang, Zihan Wang, Qing Yin, Zhengzhong Tu
A 5,049-example human-annotated dataset, reward model, and benchmark for instruction-guided video editing quality.
arXiv 2026
The Pulse of Motion: Measuring Physical Frame Rate from Visual Dynamics
Xiangbo Gao, Mingyang Wu, Siyuan Yang, Jiongze Yu, Pardis Taghavi, Fangzhou Lin, Zhengzhong Tu
Visual Chronometer estimates physical frame rate from motion and exposes temporal-scale failures in modern video generators.
arXiv 2026
PISCO: Precise Video Instance Insertion with Sparse Control
Xiangbo Gao, Renjie Li, Xinghao Chen, Yuheng Wu, Suofei Feng, Qing Yin, Zhengzhong Tu
A video diffusion model that inserts instances with coherent appearance, motion, and interaction from only a few controlled keyframes.
arXiv 2025
SafeCoop: Unravelling Full Stack Safety in Agentic Collaborative Driving
Xiangbo Gao, Tzu-Hsiang Lin, Ruojing Song, Yuheng Wu, Kuan-Ru Huang, Zicheng Jin, Fangzhou Lin, Shinan Liu, Zhengzhong Tu
An agentic defense pipeline protects language-based collaborative driving against communication failures and semantic attacks.
arXiv 2025
AirV2X: Unified Air-Ground Vehicle-to-Everything Collaboration
Xiangbo Gao, Yuheng Wu, Fengze Yang, Xuewen Luo, Keshu Wu, Xinghao Chen, Yuping Wang, Chenxi Liu, Yang Zhou, Zhengzhong Tu
A large-scale dataset and benchmark for vehicle-to-drone collaboration using flexible aerial perception.
Community
Professional Services
Conference and Journal Paper Reviewing
- CV & ML:ICCV, CVPR, ICLR, NeurIPS, T-PAMI
- Robotics:RA-L
- Transportation:TRBAM
Organization
Program Committee Member
Invited Speaker
- Invited Talk @ MEIS Workshop, CVPR 2025
- Invited Talk @ Large Language Model Applications in Civil Engineering Workshop, 2025
