Recursive paraphrasing attacks substantially lower detection rates for multiple AI text detectors with only minor quality loss, while a theoretical analysis ties best-case AUROC to total variation distance between human and AI distributions.
Protecting language generation models via invisible watermarking,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
A2-DIDM uses accumulators and ZK proofs on blockchain to verify DNN model identity from weight checkpoint sequences while protecting data and function privacy.
citing papers explorer
-
Can AI-Generated Text be Reliably Detected?
Recursive paraphrasing attacks substantially lower detection rates for multiple AI text detectors with only minor quality loss, while a theoretical analysis ties best-case AUROC to total variation distance between human and AI distributions.
-
A2-DIDM: Privacy-preserving Accumulator-enabled Auditing for Distributed Identity of DNN Model
A2-DIDM uses accumulators and ZK proofs on blockchain to verify DNN model identity from weight checkpoint sequences while protecting data and function privacy.