Jiayun (Peter) Wang

Postdoc, California Institute of Technology | peterw at caltech dot edu

prof.jpg

I am a postdoctoral researcher in the Computing + Mathematical Sciences Department at the California Institute of Technology, working with Anima Anandkumar. I am also working with Gianmarco Pinton and Lihong Wang on applying physics-aware ML for medical imaging. Prior to joining Caltech, I completed my Ph.D. in Vision Science and Berkeley AI Research (BAIR) at UC Berkeley, where I worked with Stella Yu on fundamental machine learning and computer vision models and Meng Lin on applying them to healthcare.

Research Interest: My research lies at the intersection of machine learning, computer vision and AI for healthcare. My research highlights:

  1. Minimally Supervised ML. Self-supervised learning from unlabeled data for recognition & detection (TPAMI’21) and for geometry (ECCV’24).
  2. Efficient AI Algorithms. AI models as a duality to the data. Models can be made efficient if they are aware of data structures, such as orthogonality (CVPR’20) and recurrence (WACV’23).
  3. AI for Health. Minimally supervised ML for enhanced clinical effectiveness Applications include computational imaging for ultrasound (arXiv’25) and assistive diagnosis (MICCAI’24).

Research Overview Video (Feb 2025)

Research overview

I am on the 2024-2025 job market, looking for full-time positions. Please feel free to reach out!

news

Feb 10, 2025 Congrats to my mentee Arushi for the CS PhD offer from Stanford!
Feb 09, 2025 Congrats to my mentee Aditi for the CS PhD offer from Princeton!
Dec 02, 2024 SSL for Surgery won the best paper at ML4H 2024! 🏆
Nov 25, 2024 Unified Model for MRI is on arXiv.
Aug 01, 2024 Pose-Aware Self-Supervised Learning accepted to ECCV 2024 as an oral presentation! 🎊

selected publications

  1. CVPR
    orth.png
    Orthogonal Convolutional Neural Networks
    Jiayun WangYubei ChenRudrasis Chakraborty, and Stella Yu
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020
  2. WACV
    rpg.png
    Compact and Optimal Deep Learning with Recurrent Parameter Generators
    Jiayun WangYubei ChenStella YuBrian Cheung, and Yann LeCun
    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2023
  3. ECCV
    traj.png
    Pose-Aware Self-Supervised Learning with Viewpoint Trajectory Regularization
    Jiayun WangYubei Chen, and Stella Yu
    In European Conference on Computer Vision, 2024