Targeted Structure Completion for Sparse-View 3D Reconstruction in Autonomous Driving
European Conference on Computer Vision (ECCV), 2026
About Me
I joined Ant Group as a researcher through the Ant Star Plan-A Program, where I focus on multimodal large language models with advanced agentic reasoning and interactive capabilities. I obtained Ph.D degree at Shanghai Jiao Tong University (SJTU), advised by Prof. Chao Ma. Previously, I worked as a researcher at Tencent from 2021 to 2022. I received my M.S. degree from the University of Chinese Academy of Sciences (UCAS) and the Institute of Computing Technology (ICT), Chinese Academy of Sciences, in 2021, under the supervision of Prof. Hu Han, Prof. Shiguang Shan, and Prof. Xilin Chen. I have also collaborated closely with Dr. Zhongdao Wang, Dr. Pin Tang, Dr. Xiangxuan Ren, Dr. Jilai Zheng, and Dr. Hao Lu.
Recently, my research focuses on Long-Horizon Reasoning and Interaction for Multimodal Agents. My current interests include:
- Agentic Multimodal Large Language Models
- Vison-Language-Action Models
- World Model
News
- One paper on VLA is accepted by NeurIPS 2026.
- I successfully defended my Ph.D. dissertation.
- Two papers on VLA and 3DGS are accepted by ECCV 2026.
- One paper on world model is accepted by ICME 2026 (Oral).
- Two papers on VLM and VLA are accepted by CVPR 2026.
- One paper on 3D perception is accepted by IEEE TIP.
- One paper on 3D object detection for autonomous driving is accepted by AAAI 2025.
- Two papers on 3D occupancy prediction for autonomous driving are accepted by ECCV 2024.
Earlier news (2019–2024)
- One paper on 3D occupancy prediction for autonomous driving is accepted by CVPR 2024.
- We rank first on the nuScenes occupancy prediction leaderboard (MetaGOD).
- I will pursue my Ph.D. degree at Shanghai Jiao Tong University.
- We are invited to give a talk in Valse Webinar [video].
- I will work at Tencent as an algorithm researcher.
- One paper on cross-domain face presentation attack detection (PAD) is accepted by IEEE T-IFS.
- We receive the 11th place (Top 1%) of Alibaba Tianchi Object Detection challenge.
- One paper on cross-domain face presentation attack detection (PAD) is accepted by CVPR 2020.
- One paper on multi-modal face anti-spoofing is accepted by CVPR 2019 workshop.
- One paper on cross-domain face presentation attack detection (PAD) is accepted by ICB 2019.
Publications
Publications by research area · Google Scholar ↗
Multumodal Large Language Models & World Model
PixelPilot: Scalable Vision-Language-Action Models for End-to-End Autonomous Driving
European Conference on Computer Vision (ECCV), 2026
Exploring World Transitions via 3D Occupancy World Model
IEEE International Conference on Multimedia and Expo (ICME), 2026 (Oral)
Learning Vision-Language-Action World Models for Autonomous Driving
Findings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2026
Grounding Everything in Tokens for Multimodal Large Language Models
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2026
3D Perception & Autonomous Driving
Cross-modality Feature Aggregation for Cross-domain Point Cloud Representation Learning
IEEE Transactions on Image Processing (TIP), 2026
Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering
Annual AAAI Conference on Artificial Intelligence (AAAI), 2025
OccGen: Generative Multi-modal 3D Occupancy Prediction for Autonomous Driving
European Conference on Computer Vision (ECCV), 2024
VEON: Vocabulary-Enhanced Occupancy Prediction
European Conference on Computer Vision (ECCV), 2024
SparseOcc: Rethinking Sparse Latent Representation for Vision-Based Semantic Occupancy Prediction
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024
Representation Learning & Face Anti-Spoofing
Unsupervised Adversarial Domain Adaptation for Cross-domain Face Presentation Attack Detection
IEEE Transactions on Information Forensics and Security (TIFS), 2021
Cross-domain Face Presentation Attack Detection via Multi-domain Disentangled Representation Learning
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Multi-modal Face Presentation Attack Detection via Spatial and Channel Attentions
IEEE Conference on Computer Vision and Pattern Recognition (CVPR) workshop, 2019
Improving Cross-database Face Presentation Attack Detection via Adversarial Domain Adaptation
IAPR International Conference on Biometrics (ICB), 2019
Preprints
Education
Shanghai Jiao Tong University
September 2022 – September 2026
Ph.D. in Computer Science and Technology
Department of Computer Science and Engineering
University of Chinese Academy of Sciences
August 2018 – July 2021
M.S. in Computer Technology
Institute of Computing Technology, Chinese Academy of Sciences
Experience
Ant Group · Ling Team
Researcher · September 2026 – Present · Hangzhou, China
Foundation Models, Agentic AI
V2Fun
Research Intern · June 2026 – September 2026 · Shanghai, China
Led post-training for 3D foundation models.
Huawei · Noah’s Ark Lab
Research Intern · December 2023 – April 2026 · Shanghai, China
Autonomous Driving, Generative AI, Vision-Language Models
Tencent
Researcher · July 2021 – August 2022 · Beijing, China
Video Understanding and Retrieval, 2D Perception
Peng Cheng Laboratory
Research Intern · September 2020 – January 2021 · Shenzhen, China
Domain Generalization, Face Recognition
Honors & Awards
- Tencent Outstanding Contributor (Top 10%), 2022 H1
- Tencent Good+ Contributor (Top 10%), 2021 H2
- Nomination for the Chinese Academy of Sciences President Award, 2020
- Alibaba Tianchi Object Detection Challenge — Top 1% (11th / 2,990), 2020
- CVPR 2020 Face PAD Challenge — 9th / 46, 2020
- CVPR 2019 Multi-modal Face PAD Challenge — 9th / 553, 2019
- Merit Student, UCAS, 2020
Academic Services
Conference ReviewerCVPR, ICCV, ECCV, NeurIPS, ICLR, ICML, AAAI
Journal ReviewerIEEE TIP, IEEE TIFS, IEEE TCSVT, IEEE TAI, Pattern Recognition, ACM TIST, Information Fusion