Kejiang Chen

Inspired by recent findings that self-attention acts as a low-pass filter and that its low-frequency components are informative, AttnPrint for white-box provenance is developed, which achieves strong derivative-model detection performance while remaining robust to five downstream modification techniques.

PURA achieves over three times the message match rate of existing unbiased watermarking methods, all while preserving text quality and speed.

FARI achieves faster and more robust watermark verification in diffusion models, outperforming traditional methods with a single inversion step.