A Watermark for Order-Agnostic Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Ruibo, Wu, Yihan, Chen, Yanshuo, Liu, Chenxi, Guo, Junfeng, Huang, Heng
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909353588555776
author Chen, Ruibo
Wu, Yihan
Chen, Yanshuo
Liu, Chenxi
Guo, Junfeng
Huang, Heng
author_facet Chen, Ruibo
Wu, Yihan
Chen, Yanshuo
Liu, Chenxi
Guo, Junfeng
Huang, Heng
contents Statistical watermarking techniques are well-established for sequentially decoded language models (LMs). However, these techniques cannot be directly applied to order-agnostic LMs, as the tokens in order-agnostic LMs are not generated sequentially. In this work, we introduce Pattern-mark, a pattern-based watermarking framework specifically designed for order-agnostic LMs. We develop a Markov-chain-based watermark generator that produces watermark key sequences with high-frequency key patterns. Correspondingly, we propose a statistical pattern-based detection algorithm that recovers the key sequence during detection and conducts statistical tests based on the count of high-frequency patterns. Our extensive evaluations on order-agnostic LMs, such as ProteinMPNN and CMLM, demonstrate Pattern-mark's enhanced detection efficiency, generation quality, and robustness, positioning it as a superior watermarking technique for order-agnostic LMs.
format Preprint
id arxiv_https___arxiv_org_abs_2410_13805
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Watermark for Order-Agnostic Language Models
Chen, Ruibo
Wu, Yihan
Chen, Yanshuo
Liu, Chenxi
Guo, Junfeng
Huang, Heng
Computation and Language
Statistical watermarking techniques are well-established for sequentially decoded language models (LMs). However, these techniques cannot be directly applied to order-agnostic LMs, as the tokens in order-agnostic LMs are not generated sequentially. In this work, we introduce Pattern-mark, a pattern-based watermarking framework specifically designed for order-agnostic LMs. We develop a Markov-chain-based watermark generator that produces watermark key sequences with high-frequency key patterns. Correspondingly, we propose a statistical pattern-based detection algorithm that recovers the key sequence during detection and conducts statistical tests based on the count of high-frequency patterns. Our extensive evaluations on order-agnostic LMs, such as ProteinMPNN and CMLM, demonstrate Pattern-mark's enhanced detection efficiency, generation quality, and robustness, positioning it as a superior watermarking technique for order-agnostic LMs.
title A Watermark for Order-Agnostic Language Models
topic Computation and Language
url https://arxiv.org/abs/2410.13805