Kai Xu

PhD Candidate

National University of Singapore

kxu [at] comp.nus.edu [dot] sg

[Google Scholar] [GitHub]




Biography

I am a PhD candidate (2022–present) at the School of Computing, National University of Singapore, advised by Angela Yao. As a member of the Computer Vision and Machine Learning Group, I focus on video understanding—especially 4D scene reconstruction and interpretation—to bridge visual perception and spatiotemporal reasoning in dynamic environments.

I am also a research intern at the Max Planck Institute for Intelligent Systems (MPI‑IS), advised by Michael J. Black, where my work centers on 4D reconstruction of humans and scenes from video.

Publications

  • Diagnosing Pretrained Models for Out-of-distribution Detection

    Haipeng Xiong, Kai Xu, Angela Yao

    International Conference on Computer Vision (ICCV), 2025.

  • A Constrained Optimization Approach for Gaussian Splatting from Coarsely-posed Images and Noisy Lidar Point Clouds

    Jizong Peng, Tze Ho Elden Tse, Kai Xu, Wenchao Gao, Angela Yao

    International Conference on Computer Vision (ICCV), 2025. (Highlight)

  • DAS3R: Dynamics-Aware Gaussian Splatting for Static Scene Reconstruction

    Kai Xu, Tze Ho Elden Tse, Jizong Peng, Angela Yao

    ArXiv, 2024.

  • EvA: Erasing Spurious Correlations with Activations

    Qiyuan He, Kai Xu, Angela Yao

    International Conference on Learning Representations (ICLR), 2025.

  • Enhancing Video Super-Resolution via Implicit Resampling-based Alignment

    Kai Xu, Ziwei Yu, Xin Wang, Michael Bi M, Angela Yao

    Computer Vision and Pattern Recognition (CVPR), 2024. (Highlight)

  • Scaling for Training Time and Post-hoc Out-of-distribution Detection Enhancement

    Kai Xu, Rongyu Chen, Gianni Franchi, Angela Yao

    International Conference on Learning Representations (ICLR), 2024.

  • DropIT: Dropping Intermediate Tensors for Memory-Efficient DNN Training

    Joya Chen*, Kai Xu*, Yuhui Wang, Yifei Cheng, Angela Yao (*Equal contribution)

    International Conference on Learning Representations (ICLR), 2023.

  • Accelerating Video Object Segmentation with Compressed Video

    Kai Xu, Angela Yao

    Computer Vision and Pattern Recognition (CVPR), 2022.

    • Contact

    Address: AS6 04-11, 11 Computing Dr, Singapore 117416