Ze Yang    杨泽

I am a Ph.D. candidate at the University of Toronto and the Vector Institute, advised by Prof. Raquel Urtasun. I am also a Staff Research Scientist and Tech Lead at Waabi, focusing on self-driving technology. Before joining Waabi, I was a Research Scientist at Uber ATG. Prior to moving to Toronto, I received a Master's Degree in Computer Science from Peking University, advised by Prof. Liwei Wang.

My research lies at the intersection of 3D computer vision, robotics, and machine learning. I am building scalable world modeling and simulation systems that facilitate the safe, controlled, and cost-effective development of autonomous systems.

Email  /  CV  /  Google Scholar  /  Github  /  LinkedIn  /  Twitter

• ^NEW 2025/03: GenAssets (in-the-wild 3D assets reconstruction and generation) was accepted by CVPR 2025.
• 2024/07: Our work on Neural Sensor Calibration and Generalizable Reconstruction were accepted by ECCV 2024.
• 2024/01: Our work on 4D World Model for Self-Driving was accepted by ICLR 2024. Congrats to Lunjun!
• 2023/12: I gave a talk at MILA Robot Learning Seminar on Learning in-the-wild Sensor Simulation for Autonomous Driving.
• 2023/09: LightSim (lighting simulation) was accepted by NeurIPS 2023. Congrats to Ava, Gary and Jingkang!
• 2023/07: I gave a talk at Toronto Computational Imaging Group on in-the-wild 3D reconstruction, modeling and sensor simulation.
• 2023/07: NeuRas (large scene reconstruction & rendering) was accepted by ICCV 2023. Congrats to Jeffrey!
• 2023/07: LiDAR-DG (simulation domain gap analysis) was accepted by ICCV 2023. Congrats to Siva and Andrei!
• 2023/03: UniSim (3D scene reconstruction, modeling and closed-loop sensor simulation for self-driving) was accepted by CVPR 2023 as highlight.
• 2023/01: NeuSim (3D object reconstruction for sensor simulation) was accepted by ICRA 2023.
• 2022/10: CADSim (3D object reconstruction for sensor simulation) was accepted by CoRL 2022.
• 2022/10: I gave a talk at ByteDance Research on in-the-wild 3D reconstruction for sensor simulation.
• 2022/03: RBGNet (3D object detection) was accepted by CVPR 2022. Congrats to Haiyang!
• 2021/06: I gave a talk at CVPR 2021 tutorial All About Self-Driving on dynamic asset reconstruction for simulation.
• 2021/03: I joined Waabi to work on next-generation self-driving solution.
• 2021/02: S3 (3D human reconstruction and modeling) was accepted by CVPR 2021.
• 2020/10: LiME (3D human reconstruction, modeling and sensor simulation) was accepted by CoRL 2020.
• 2020/09: I started my Ph.D. journey at the University of Toronro.
• 2020/06: I joined Uber Advanced Technologies Group (Uber ATG) to work on self-driving technology.
• 2020/06: I graduated from Peking University.
• 2020/06: Follow-up work Dense RepPoints was accepted by ECCV 2020.
• 2019/07: RepPoints (representing objects as point sets for visual perception) was accepted by ICCV 2019.
• 2018/12: I started my internship at Microsoft Research Asia to work on visual perception.

My research goal is to build scalable and realistic digital twins for real-world modeling, with the purpose of creating immersive and controllable world simulations that facilitate the development and evaluation of autonomous systems, such as self-driving vehicles, in a safe, controlled, reactive, and cost-effective manner. Towards this goal, I have delved into various areas over the past few years, such as reconstruction spanning from individual objects to large-scale scene; world modeling encompassing both rigid and dynamic content; and closed-loop simulation for camera and LiDAR data. During the earlier stages of my research, I'm interested in learning flexible and structural representation for visual perception.

Waabi

Uber Advanced Technologies Group

Microsoft Research Asia

• Conference Reviewer: CVPR, ICCV, ECCV, ACCV, WACV, ICRA, IROS, NeurIPS, ICLR, ICML, AAAI


• Journal Reviewer: TPAMI, TCSVT, TMM

Website template comes from Jon Barron.