PeopleofSuperintelligence

Researcher

Peng Dai

Peng Dai is a computer vision researcher who works on neural rendering, video generation and 3D content creation. At the University of Electronic Science and Technology of China, he wrote Neural Point Cloud Rendering via Multi-Plane Projection, published at CVPR 2020. With Xiaojuan Qi, who supervised his Ph.D. at the University of Hong Kong, he coauthored papers on video demoireing at CVPR 2022, hybrid neural rendering at CVPR 2023, and SVG, a method for 3D stereoscopic video generation presented at ICLR 2025. He also worked at Google as a student researcher, hosted by Yinda Zhang, on immersive video and 4D content generation. He is a Distinguished Postdoctoral Fellow at the Vector Institute, hosted by Renjie Liao, and studies world models, 3D/4D vision and embodied AI.

Timeline

  1. 2020

    Neural point cloud rendering

    At the University of Electronic Science and Technology of China, he writes Neural Point Cloud Rendering via Multi-Plane Projection with Yinda Zhang, Zhuwen Li, Shuaicheng Liu and Bing Zeng, published in the CVPR 2020 proceedings. The pipeline projects point-cloud features into a layered camera-frustum volume and blends the resulting depth layers into novel-view images.

  2. 2022

    Demoireing video and images

    In 2022, he coauthors Video Demoireing with Relation-Based Temporal Consistency at CVPR, which introduces the first hand-held video demoireing dataset. He also coauthors, with Xin Yu and Xiaojuan Qi, an ECCV 2022 paper on ultra-high-definition image demoireing.

  3. 2023

    Hybrid rendering with motion blur

    In 2023, he coauthors Hybrid Neural Rendering for Large-Scale Scenes With Motion Blur at CVPR with Yinda Zhang, Xin Yu, Xiaoyang Lyu and Xiaojuan Qi. The model joins image-based and neural 3D representations to render view-consistent images from in-the-wild photos.

  4. 2024–2025

    Stereoscopic video generation

    In June 2024, he posts SVG: 3D Stereoscopic Video Generation via Denoising Frame Matrix on arXiv with Feitong Tan, Qiangeng Xu, Yinda Zhang, Xiaojuan Qi and others, and the paper appears at ICLR in 2025. The method turns the output of an off-the-shelf monocular video generator into 3D stereoscopic video without scene optimization or model fine-tuning.

  5. 2025

    GO-NeRF for VR content

    In 2025, IEEE Transactions on Visualization and Computer Graphics publishes GO-NeRF: Generating Objects in Neural Radiance Fields for Virtual Reality Content Creation, written with Feitong Tan, Xin Yu, Yifan Peng, Yinda Zhang and Xiaojuan Qi. The paper also appears at IEEE VR.

  6. 2026

    S2VG in IEEE TPAMI

    In 2026, IEEE Transactions on Pattern Analysis and Machine Intelligence publishes S2VG: 3D Stereoscopic and Spatial Video Generation via Denoising Frame Matrix, which he writes with Feitong Tan, Qiangeng Xu, Yihua Huang, David Futschik, Ruofei Du, Sean Fanello, Yinda Zhang and Xiaojuan Qi.

  7. As of October 2026

    Postdoctoral fellow at Vector

    He is a Distinguished Postdoctoral Fellow at the Vector Institute, hosted by Prof. Renjie Liao, after a Ph.D. at the University of Hong Kong under Prof. Xiaojuan Qi. He previously worked at Google as a student researcher, hosted by Dr. Yinda Zhang, on immersive video and 4D content generation. His research interests include world models, 3D/4D vision and embodied AI.

Career

  1. NowVector InstituteDistinguished Postdoctoral Fellow
  2. GoogleStudent Researcher

Education

  1. University of Hong KongPhD

From the biography.

  1. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes Peng Dai. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.
  2. Nonprofit roles are 2.3 times as common in AI safetyIncludes Vector Institute. We estimate where people focused on safety, alignment and evaluation work. Nonprofit roles appear for 25.1% of safety people with known affiliations, compared with 11.0% in other areas.
  3. Universities (31%) and startups and other companies (25%) employ the largest shares of 2020–2026 entrantsIncludes Vector Institute. Among 819 people whose first recorded role started in 2020–2026 and who have a known current organization, 60% work in commercial sectors (25% at startups and other companies, 23% at frontier labs and 13% at big tech, with 9 in two), and 31% work at universities, compared with 66% for earlier entrants.
  4. The United States employs an estimated 48% of people shaping AIIncludes Vector Institute. We estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Peng Dai yet. Send us a free-licensed photo.

  1. [1912.04645] Neural Point Cloud Rendering via Multi-Plane Projection ar5iv.labs.arxiv.org. Checked 22 October 2026.
  2. Neural Point Cloud Rendering via Multi-Plane Projection | ML Anthology mlanthology.org. Checked 22 October 2026.
  3. Peng Dai - Google Scholar scholar.google.com. Checked 22 October 2026.
  4. Xiaojuan Qi xjqi.github.io. Checked 22 October 2026.
  5. [2204.02957] Video Demoireing with Relation-Based Temporal Consistency arxiv.org. Checked 22 October 2026.
  6. CVPR 2023 Open Access Repository openaccess.thecvf.com. Checked 22 October 2026.
  7. [2407.00367] SVG: 3D Stereoscopic Video Generation via Denoising Frame Matrix arxiv.org. Checked 22 October 2026.
  8. Peng Dai - Vector Institute for Artificial Intelligence vectorinstitute.ai. Checked 22 October 2026.