Avatar

Zeyu Wang

PhD in Computer Science and Engineering

UC Santa Cruz

I am a 3rd-year PhD student at UC Santa Cruz, supervised by Prof. Cihang Xie. Before this, I received my M.S. degree at Institute of Computing Technology, Chinese Academy of Science in 2021, and my B.E. degree at Central South University in 2018. I am an incoming intern at LLNL. have also spent wonderful times at QCraft and Megvii (Face++).

My research interest lies in the intersection of computer vision and machine learning.

Education

  • PhD in Computer Science and Engineering, 2021 - Now

    UC Santa Cruz

  • MS in Computer Science, 2018 - 2021

    Institute of Computing Technology, Chinese Academy of Science

  • BE in Automation, 2014 - 2018

    Central South University

Publications

Revisiting Adversarial Training at Scale

Rejuvenating image-GPT as Strong Visual Representation Learners

Sculpting Holistic 3D Representation in Contrastive Language-Image-3D Pre-training

FedConv: Enhancing Convolutional Neural Networks for Handling Data Heterogeneity in Federated Learning

DistillBEV: Boosting Multi-Camera 3D Object Detection with Cross-Modal Knowledge Distillation

An Inverse Scaling Law for CLIP Training

On the Adversarial Robustness of Camera-based 3D Object Detection

Masked Autoencoders Enable Efficient Knowledge Distillers

Can CNNs Be More Robust Than Transformers?

Siamfc++: Towards robust and accurate visual tracking with target estimation guidelines

Services

Teaching Assistant

  • UC Santa Cruz - CSE164 (Computer Vision), Spring 2022
  • UC Santa Cruz - California State Summer School for Mathematics and Science (COSMOS), Summer 2023
  • UC Santa Cruz - CSE144 (Applied Machine Learning), Fall 2023

Reviewer

  • Conference reviewer: AAAI 2022, NeuRIPS 2022-2024, ICLR2023-2024, CVPR2023-2024, ICML2023-2024, ICCV2023, WACV2024
  • Journal Reviewer:
    • IEEE Transactions on Pattern Analysis and Machine Intelligence
    • International Journal of Computer Vision
    • IEEE Transactions on Circuits and Systems for Video Technology