Kaixiang Zheng 郑凯翔

Ph.D. Candidate @ University of Waterloo

I am a Ph.D. candidate in the Department of Electrical and Computer Engineering at the University of Waterloo, advised by Prof. En-Hui Yang in the Multicom Research Group. Before that, I received my B.S. in Automation from Tianjin University in 2021.

My research mainly focuses on efficient and trustworthy AI. I study how to make deep learning models more efficient through distillation, quantization and architecture design, more reliable by improving their generalization, robustness and ability to abstain. I'm also broadly interested in computer vision, information theory, large language models, and multimodality.

Kaixiang Zheng

News

Post-hoc Selective Classification for Reliable Synthetic Image Detection is accepted to NeurIPS 2026. This work was done during my internship at Reality Defender.

I presented Deep Learning Based on Train-Test Discrepancy Minimization at the VII AMMCS International Conference.

I joined Reality Defender as a Computer Vision Ph.D. Intern, working on reliable synthetic image detection.

Selected Publications

Google Scholar

* Equal contribution.

Teaching

Teaching Assistant · University of Waterloo

  • ECE 307 · Probability Theory and Statistics 2Fall 2026
  • ECE 203 · Probability Theory and Statistics 1Spring 2025
  • ECE 457B · Fundamentals of Computational IntelligenceWinter 2024
  • ECE 380 · Analog Control SystemsWinter & Spring 2023

Academic Service

Reviewer

  • ICLR2025, 2026, 2027
  • NeurIPS2026
  • CVPR2025
  • ICCV2025

Get in touch

Feel free to reach out about research and collaboration.

k56zheng@uwaterloo.ca