Zunnan Xu

Student at Tsinghua University

Hi, I am a student at Tsinghua University. Currently, my research centers on multi-modal learning and generative models, with the ultimate goal of building a unified perception–generation engine that can understand, predict, and interact with the physical world.

I am especially interested in how these models can be distilled into efficient engines that enable systems to perform robust visual reasoning and generative modeling in complex, unstructured environments. I now focus more on rethinking problems and offering simple, effective solutions. If you have any use cases you would like to share, please feel free to contact me!

News

2025	One paper accepted to NeurIPS 2025.
2025	Two papers accepted to ICCV 2025.
2025	Received Outstanding Reviewer Award at CVPR 2025.
2025	Three papers accepted to CVPR 2025.
2024	One paper accepted to EMNLP 2024.
2024	Won 3rd prize in CVPR 2024 OVD challenge.
2024	Two papers accepted to ICML 2024 workshop.
2024	One paper accepted to NeurIPS 2024.
2023	One paper accepted to ICCV 2023.

Research Interests

Generative Vision & World Models: Scaling Diffusion Transformers (DiT) and Flow Matching to build high-fidelity physics simulators (e.g., HunyuanVideo) that serve as the "brain's imagination".
Embodied Foundation Models: Integrating world modeling into unified VLA architectures (e.g., WALL-OSS) to achieve long-horizon manipulation and generalization.
Multimodal Understanding: Leveraging parameter-efficient fine-tuning (PEFT) and reinforcement learning (RL) to align vision and language features and enable fine-grained multi-modal understanding.

Selected Publications

(* denotes equal contribution)

2D Video & Image Generation

HunyuanPortrait: Implicit condition control for enhanced portrait animation

Zunnan Xu, Zhentao Yu, Zixiang Zhou, Jun Zhou, and others

CVPR 2025

HunyuanVideo: A Systematic Framework For Large Video Generative Models

Zunnan Xu (Core contributor) and Hunyuan Foundation Model Team

Technical Report, 2024

Audio-visual controlled video diffusion with masked selective state spaces modeling for natural talking head generation

Fa-Ting Hong, Zunnan Xu, Zixiang Zhou, Jun Zhou, Xiu Li, Qin Lin, Qinglin Lu, Dan Xu

ICCV 2025

Alignment is All You Need: A Training-free Augmentation Strategy for Pose-guided Video Generation

Xiaoyu Jin*, Zunnan Xu*, Mingwen Ou, Wenming Yang

CVG@ICML 2024

Zero-shot 3D-Aware Trajectory-Guided image-to-video generation via Test-Time Training

Ruicheng Zhang*, Jun Zhou*, Zunnan Xu*, Zihao Liu, Jiehui Huang, Mingyang Zhang, Yu Sun, Xiu Li

AAAI 2026

Fireedit: Fine-grained instruction-based image editing via region-aware vision language model

Jun Zhou, Jiahao Li, Zunnan Xu, Hanhui Li, Yiji Cheng, Fa-Ting Hong, Qin Lin, Qinglin Lu, Xiaodan Liang

CVPR 2025

InterAnimate: Taming Region-aware Diffusion Model for Realistic Human Interaction Animation

Yukang Lin, Yan Hong, Zunnan Xu, Xindi Li, Chao Xu, Chuanbiao Song, Ronghui Li, Haoxing Chen, Jun Lan, Huijia Zhu, and others

ACMMM 2025

Vision-Language Models

Bridging vision and language encoders: Parameter-efficient tuning for referring image segmentation

Zunnan Xu, Zhihong Chen, Yong Zhang, Yibing Song, Xiang Wan, Guanbin Li

ICCV 2023

Igniting vlms toward the embodied space

Zunnan Xu (Core contributor) and X Square Robot Team

Technical Report, 2025

SAM-R1: Leveraging SAM for Reward Feedback in Multimodal Segmentation via Reinforcement Learning

Jiaqi Huang*, Zunnan Xu*, Jun Zhou, Ting Liu, Yicheng Xiao, Mingwen Ou, Bowen Ji, Xiu Li, Kehong Yuan

NeurIPS 2025

Enhancing Fine-grained Multi-modal Alignment via Adapters: A Parameter-Efficient Training Framework

Zunnan Xu, Jiaqi Huang, Ting Liu, Yong Liu, Haonan Han, Kehong Yuan, Xiu Li

WANT@ICML 2024

Densely Connected Parameter-Efficient Tuning for Referring Image Segmentation

Jiaqi Huang*, Zunnan Xu*, Ting Liu, Yong Liu, Haonan Han, Kehong Yuan, Xiu Li

AAAI 2025

MaPPER: Multimodal Prior-guided Parameter Efficient Tuning for Referring Expression Comprehension

Ting Liu*, Zunnan Xu*, Yue Hu, Liangtao Shi, Zhiqiang Wang, Quanjun Yin

EMNLP 2024

3D Motion Synthesis

Mambatalk: Efficient holistic gesture synthesis with selective state space models

Zunnan Xu, Yukang Lin, Haonan Han, Sicheng Yang, Ronghui Li, Yachao Zhang, Xiu Li

NeurIPS 2024

Chain of generation: Multi-modal gesture synthesis via cascaded conditional control

Zunnan Xu, Yachao Zhang, Sicheng Yang, Ronghui Li, Xiu Li

AAAI 2024

Separate to Collaborate: Dual-Stream Diffusion Model for Coordinated Piano Hand Motion Synthesis

Zihao Liu*, Mingwen Ou*, Zunnan Xu*, Jiaqi Huang, Haonan Han, Ronghui Li, Xiu Li

ACMMM 2025

Freetalker: Controllable speech and text-driven gesture generation based on diffusion models

Sicheng Yang, Zunnan Xu, Haiwei Xue, Yongkang Cheng, Shaoli Huang, Mingming Gong, Zhiyong Wu

ICASSP 2024

Atom: Aligning text-to-motion model at event-level with gpt-4vision reward

Haonan Han, Xiangzuo Wu, Huan Liao, Zunnan Xu, Zhongyuan Hu, Ronghui Li, Yachao Zhang, Xiu Li

CVPR 2025

3D Generation

Reparo: Compositional 3d assets generation with differentiable 3d layout alignment

Haonan Han, Rui Yang, Huan Liao, Jiankai Xing, Zunnan Xu, Xiaoming Yu, Junwei Zha, Xiu Li, Wanhua Li

ICCV 2025

Consistent123: One image to highly consistent 3d asset using case-aware diffusion priors

Yukang Lin, Haonan Han, Chaoqun Gong, Zunnan Xu, Yachao Zhang, Xiu Li

ACMMM 2024

Selected Preprint

MIND-V: Hierarchical Video Generation for Long-Horizon Robotic Manipulation with RL-based Physical Alignment

Zunnan Xu (Project Lead)

Under Review

Honors & Awards

2025	Outstanding Fellowship of Tencent Rhino-Bird Research Elite Program
2025	Golden Award in International Exhibition of Inventions Geneva
2024	Graduate National Scholarship
2024	Tencent Rhino-Bird Research Elite Program Student
2023	Outstanding Graduate (Bachelor's Level)
2022	Golden Award in iGEM Competition
2021	Golden Award in National College Student Algorithm Design and Programming Challenge
2020	Undergraduate National Scholarship

Academic Services

Reviewer for:

Journal: IEEE TPAMI, IEEE TIP, IJCV, TMLR, MIR, AI
Conference: CVPR, ICCV, NeurIPS, ICML, ICLR, ACL, EMNLP, NAACL, WACV