Xiaoyu Wu

Xiaoyu Wu

PhD student working on privacy, trustworthy AI, and generative models.

Adversarial Example Does Good: Preventing Painting Imitation from Diffusion Models via Adversarial Examples

Published in ICML 2023 (Oral), 2023

This work develops a theoretical framework for adversarial examples against diffusion models and introduces AdvDM, a Monte Carlo method that produces protective watermarks for human-created artwork.

Recommended citation: Liang, C.*, Wu, X.*, Hua, Y., et al. (2023). "Adversarial Example Does Good: Preventing Painting Imitation from Diffusion Models via Adversarial Examples." ICML 2023. Oral presentation.
Download Paper