Xin Zhou | HUST

About Me

I am a second-year Ph.D. student at Huazhong University of Science and Technology, under the supervision of Prof. Xiang Bai. My research interests mainly lie in 3D vision, especially in 3D understanding and generative models.

Research interests: foundation models, vision-language-action (VLA) models, world models, and world action models (WAM) for autonomous driving and embodied AI.

News

Scroll for earlier updates

[Sep. 2026] Our work Qwen-Drive-1.0, a vision-language foundation model for autonomous driving, is released.
[Sep. 2026] HERMES++ is accepted to TPAMI.
[Jul. 2026] Selected as an Outstanding Reviewer for ECCV 2026.
[Jun. 2026] One paper is accepted to ECCV 2026.
[May. 2026] Recognized as a Silver Reviewer for ICML 2026.
[Apr. 2026] Received an Excellence in Reviewing Award from Pattern Recognition.
[Feb. 2026] Two papers are accepted to CVPR 2026.
[Oct. 2025] Selected as the Top Reviewer at NeurIPS 2025.
[Sep. 2025] One paper is accepted to NeurIPS 2025.
[Jul. 2025] One paper is accepted to TPAMI 2025.
[Jul. 2025] We release a training-free diffusion acceleration framework, now natively integrated into ComfyUI.
[Jun. 2025] Finally achieve the bachelor's degree.
[Jun. 2025] HERMES is accepted to ICCV 2025.
[Apr. 2025] We win the 2-nd place in the SoccerNet Monocular Depth Estimation Challenges (CVPR 2025).
[Feb. 2025] One paper is accepted to CVPR 2025.
[Sep. 2024] Two papers are accepted to NeurIPS 2024.
[Sep. 2024] We win the 1-st place in the FishNet Classification Challenge (ECCV 2024).
[Apr. 2024] Invited to give a talk at FALML 2024.
[Feb. 2024] One paper is accepted to CVPR 2024!
[Jan. 2024] I will keep updating a collection of World Models 2,254 — check it out and give a star🌟!
[Aug. 2023] One paper is accepted to PRCV 2023.

Selected Publications

Full list

* Equal contribution

Vision(-Language)(-Action)

Vision-language(-action) foundation model pretraining for autonomous driving and embodied AI, plus unified generation and understanding.

arXiv'26

Qwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving

Xin Zhou, Zongchuang Zhao, Zhibo Yang, Mingsheng Li, Humen Zhong, Shuai Bai, Du Chu, Ruizhe Chen, Zhaohai Li, Jun Tang, Qiuyue Wang, Mingkun Yang, Jiazhao Zhang, Dayiheng Liu, Dingkang Liang, Xiang Bai

PDF Code Blog Star 383

A vision-language foundation model for autonomous driving that unifies 3D perception, driving VQA, and motion planning.
TPAMI'26/ICCV'25

HERMES++: Toward a Unified Driving World Model for 3D Scene Understanding and Generation

ICCV 2025 version HERMES: A Unified Self-Driving World Model for Simultaneous 3D Scene Understanding and Generation

Xin Zhou, Dingkang Liang, Sifan Tu, Xiwu Chen, Yikang Ding, Dingyuan Zhang, Feiyang Tan, Hengshuang Zhao, Xiang Bai

TPAMI PDF ICCV PDF Code Project 263+71

A unified driving world model enabling joint multi-view understanding and future lidar generation.

arXiv'26 Qwen-VLA: Unifying Vision-Language-Action Modeling Across Tasks, Environments, and Robot Embodiments A unified vision-language-action model spanning tasks, environments, and robot embodiments.
TIP'26 Extending Large Vision-Language Model for Diverse Interactive Tasks in Autonomous Driving Zongchuang Zhao, Haoyu Fu, Dingkang Liang, Xin Zhou, Dingyuan Zhang, Hongwei Xie, Bing Wang, Xiang Bai Extends a large VLM from perception-level QA to diverse interactive driving tasks.

Generative World Model

Driving and general-purpose world models, world action models, and efficient video generation.

arXiv'25

Less is Enough: Training-Free Video Diffusion Acceleration via Runtime-Adaptive Caching

Xin Zhou, Dingkang Liang, Kaijin Chen, Tianrui Feng, Xiwu Chen, Hongkai Lin, Yikang Ding, Feiyang Tan, Hengshuang Zhao, Xiang Bai

PDF Project Code Star 299

A training-free runtime-adaptive caching scheme that accelerates video diffusion; officially integrated into ComfyUI, plug-and-play for the Wan series, MiniMax H3, and more.

arXiv'26 SimWAM: A Simple World Action Model for End-to-End Autonomous Driving Zongchuang Zhao, Xin Zhou, Tianyang Xu, Zhengyang Sun, Kaixuan Zhou, Yu Wu, Honglin Li, Dingkang Liang, Xiang Bai A simple world action model for end-to-end autonomous driving.
ECCV'26 Divide and Conquer: Decoupled Representation Alignment for Multimodal World Models Junyuan Xiao, Dingkang Liang, Xin Zhou, Yixuan Ye, Tongtong Su, Guangmo Yi, Bin Xia, Qiang Lyu, Shurui Shi, Jun Huang, Jianlou Si, Wenming Yang Decoupled representation alignment for multimodal world models.
CVPR'26 When Numbers Speak: Aligning Textual Numerals and Visual Instances in Text-to-Video Diffusion Models Zhengyang Sun, Yu Chen, Xin Zhou, Xiaofan Li, Xiwu Chen, Dingkang Liang, Xiang Bai Aligns textual numerals with visual instances in text-to-video diffusion.
arXiv'26 Out of Sight but Not Out of Mind: Hybrid Memory for Dynamic Video World Models Kaijin Chen, Dingkang Liang, Xin Zhou, Yikang Ding, Xiaoqiang Liu, Pengfei Wan, Xiang Bai A hybrid memory mechanism for dynamic video world models.
ICRA'26 UniFuture: A 4D Driving World Model for Future Generation and Perception Dingkang Liang, Dingyuan Zhang, Xin Zhou, Sifan Tu, Tianrui Feng, Xiaofan Li, Yumeng Zhang, Mingyang Du, Xiao Tan, Xiang Bai A unified driving world model for future generation and future perception.
FCS'26
The Role of World Models in Shaping Autonomous Driving: A Comprehensive Survey Survey
Sifan Tu, Xin Zhou, Dingkang Liang, Xingyu Jiang, Yumeng Zhang, Xiaofan Li, Xiang Bai
Collection Star 2,256
A comprehensive survey of world models for autonomous driving.

3D Perception

Point cloud and 3D scene understanding.

TPAMI'25

Parameter-Efficient Fine-Tuning in Spectral Domain for Point Cloud Learning

Dingkang Liang*, Tianrui Feng*, Xin Zhou*, Yumeng Zhang, Zhikang Zou, Xiang Bai

PDF Code Star 389

A spectral domain perspective for point PEFT, very strong performance.
CVPR'24

Dynamic Adapter Meets Prompt Tuning: Parameter-Efficient Transfer Learning for Point Cloud Analysis

Xin Zhou*, Dingkang Liang*, Wei Xu, Xingkui Zhu, Yihan Xu, Zhikang Zou, Xiang Bai

PDF Code Slide Blog Star 171

Introduce Dynamic Adapter with prompt tuning in point PEFT.
NeurIPS'24

PointMamba: A Simple State Space Model for Point Cloud Analysis

Dingkang Liang*, Xin Zhou*, Wei Xu, Xingkui Zhu, Zhikang Zou, Xiaoqing Ye, Xiao Tan, Xiang Bai

PDF Project Code Blog Star 548

A simple state space model tailored for point cloud analysis.

CVPR'26 PointTPA: Dynamic Network Parameter Adaptation for 3D Scene Understanding Siyuan Liu, Chaoqun Zheng, Xin Zhou, Tianrui Feng, Dingkang Liang, Xiang Bai Dynamic test-time parameter adaptation for 3D scene understanding.
NeurIPS'24 A Unified Framework for 3D Scene Understanding Wei Xu, Chunsheng Shi, Sifan Tu, Xin Zhou, Dingkang Liang, Xiang Bai A unified framework for multi-task 3D scene understanding.
PRCV'23 Diffusion-based 3D Object Detection with Random Boxes Xin Zhou, Jinghua Hou, Tingting Yao, Dingkang Liang, Zhe Liu, Zhikang Zou, Xiaoqing Ye, Jianwei Cheng, Xiang Bai Diffusion-based 3D object detection trained with randomly sampled boxes.

Selected Experience

Participated Projects

National Undergraduate Training Projects for Innovation and Entrepreneurship
- 2023-2024, 20,000￥ (~2,750 $). (As the host; Rated Excellent)
- 2024-2025, 10,000￥ (~1,375 $). (As a member; Rated Excellent)
Taihu Lake Innovation Fund for Future Technology (200,000￥, ~27,500 $), 2023-2025. (As a student member)
HUST Undergraduate Natural Science Innovation Fund Program, (50,000￥, ~6,900 $), 2024-2025. (As the host)

Academic Service

Conference Reviewer: CVPR, ECCV, ICML, AAAI, WACV, ICCV, ICLR, NeurIPS, ICRA, ACCV
Journal Reviewer: TPAMI, IJCV, SCIS, PR, RA-L, IJDE, TNNLS, TMLR