notesum.ai

Published at April 25

GuardT2I: Defending Text-to-Image Models from Adversarial Prompts

NeurIPS

Released Date: April 25, 2024

Authors: Yijun Yang¹, Ruiyuan Gao¹, Xiao Yang², Jianyuan Zhong¹, Qiang Xu¹

Aff.: ¹The Chinese University of Hong Kong; ²Tsinghua University

Arxiv: https://openreview.net/pdf/a5df915bb3f84deba44b7362dd79edb743d577f7.pdf