Zhao
A scholar in the security research field who has studied the limitations of black-box query attacks and randomization defenses, pointing out that random noise is averaged out statistically as the query budget grows.
Timeline (last 90 days)
Zhao et al. repeatedly verified in their research that when the query budget is scaled up to hundreds of queries or more, the gradient estimation error introduced by randomized defenses decays rapidly.
A commentary article published on arXiv (arXiv:2609.15996) challenges Chen, Zhao, and Cohan's distributional evaluation of research ideas generated by large language models.
All Facts (2)
Zhao et al. repeatedly verified in their research that when the query budget is scaled up to hundreds of queries or more, the gradient estimation error introduced by randomized defenses decays rapidly.
50%UnverifiedA commentary article published on arXiv (arXiv:2609.15996) challenges Chen, Zhao, and Cohan's distributional evaluation of research ideas generated by large language models.
50%