Establishes an unconditional robustness threshold of 1-1/q for zero-bit tamper-detection codes in watermarking, with matching constructions and experimental confirmation on image models.
Robust distortion-free watermarks for language models
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
representative citing papers
Establishes sample complexity bounds for watermark proportion estimation under Gumbel-max, showing full observation is more efficient than pivotal reduction.
citing papers explorer
-
The Coding Limits of Robust Watermarking for Generative Models
Establishes an unconditional robustness threshold of 1-1/q for zero-bit tamper-detection codes in watermarking, with matching constructions and experimental confirmation on image models.
-
Sample Complexities of Estimating Gumbel--Max Watermark Proportions with and without Reduction to Pivotal Statistics
Establishes sample complexity bounds for watermark proportion estimation under Gumbel-max, showing full observation is more efficient than pivotal reduction.