Zesun Yang

Papers

Total Citations

H-Index

About

Zesun Yang is a researcher whose work sits at the intersection of cybersecurity, adversarial machine learning, and human-computer interaction. Yang is best known for a provocative 2018 study, "Breaking Google reCaptcha V2," which demonstrated a practical method to defeat one of the web’s most widely used bot-detection systems. By training two distinct neural network classifiers to distinguish road signs from background imagery, Yang achieved high accuracy in solving reCaptcha’s visual puzzles—exposing a critical vulnerability in the CAPTCHA paradigm. This work, cited over 2,000 times, has become a foundational reference in the field of adversarial CAPTCHA breaking, influencing subsequent research on both attack and defense strategies. Beyond this landmark paper, Yang’s broader contributions include advancing understanding of how deep learning models can be exploited to bypass security measures, as well as exploring the ethical implications of such techniques. Yang’s research continues to shape conversations around the robustness of human verification systems, making their work essential reading for anyone interested in the evolving arms race between AI-driven attacks and defensive security.