Sandcastles in the Storm: Revisiting the (Im)possibility of Strong Watermarking (2025.acl-long)
Copied to clipboard
Fabrice Y Harel-Canada, Boran Erol, Connor Choi, Jason Liu, Gary Jiarui Song, Nanyun Peng, Amit Sahai
| Challenge: | Recent theoretical work argues that any watermark can be erased via random walk attacks that perturb text while preserving quality. |
| Approach: | They propose that any watermark can be erased via random walk attacks that perturb text while preserving quality. |
| Outcome: | The proposed method underperforms the theoretical models in large-scale experiments and human-validated assessments. |
Similar Papers
Revisiting the Robustness of Watermarking to Paraphrasing Attacks (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent watermarking techniques slightly modify the output probabilities of LMs to embed a signal in the generated output that can later be detected. |
| Approach: | They propose to use a model to produce a watermarking signal that is invariant to semantically-similar inputs to undo the effects of watermarks. |
| Outcome: | The proposed method undoes the effects of watermarking and dramatically improves the effectiveness of paraphrasing attacks with limited access to model generations. |
Watermark Smoothing Attacks against Language Models (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Watermarking is a key technique for detecting AI-generated text. |
| Approach: | They propose a method to selectively smooth watermarks by leveraging the relationship between the model’s confidence and detectability. |
| Outcome: | The proposed method selectively smoothes watermark traces while preserving text quality. |
Subtle Signatures, Strong Shields: Advancing Robust and Imperceptible Watermarking in Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have led to an increase in AI-generated text on the Internet, presenting a crucial challenge to differentiate AI-created content from human-written text. |
| Approach: | They propose a novel approach to embed watermarks into LLMs that leverages token prior probabilities to improve detectability and maintain watermark imperceptibility. |
| Outcome: | The proposed method improves detectability and imperceptibility of watermarks by partitioning tokens into two distinct groups based on prior probabilities and employing tailored strategies for each group. |
Watermark under Fire: A Robustness Evaluation of LLM Watermarking (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Various watermarking methods have been proposed to identify LLM-generated texts . lack of unified evaluation platforms has left many critical questions unanswered . |
| Approach: | They systematize existing LLM watermarkers and watermark removal attacks and develop a unified platform that integrates them. |
| Outcome: | The proposed systematizes existing LLM watermarkers and watermark removal attacks, mapping out their design spaces. |
Synthetic Text Detection in the Age of Large Language Models: Watermark vs. Automatic Detection (2026.acl-industry)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are ubiquitous and capable of generating long coherent texts that look almost indistinguishable from human-written texts. |
| Approach: | They propose to use watermark and automatic detection to detect synthetic texts generated from Large Language Models (LLMs) they evaluate six different models, six different watermark techniques and two different automatic detectors for different levels of syntactic changes. |
| Outcome: | The proposed methods outperform on unperturbed and perturbed datasets on six different sizes of Qwen2.5 models, six watermark techniques and detectors, and two automatic detectors. |
Beyond A Fixed Seal: Adaptive Stealing Watermark in Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing detection methods for large language models rely on fixed strategies to steal watermarks. |
| Approach: | They propose a novel steal-based watermark algorithm that derives watermark information from watermarked texts to craft highly targeted adversarial attacks. |
| Outcome: | The proposed system significantly increases steal efficiency against target watermarks under identical conditions. |
MorphMark: Flexible Adaptive Watermarking for Large Language Models (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for tracing text origins struggle with a watermark effectiveness dilemma . weaker watermarks preserve text quality, while stronger ones enhance effectiveness . |
| Approach: | They propose a method that adjusts watermark strength in response to changes in a key factor . they first formalize the problem within a multi-objective trade-off analysis framework . |
| Outcome: | The proposed method improves watermark effectiveness but reduces text quality . the proposed method prioritizes flexibility and time and space efficiency . |
Paraphrasing Attack Resilience of Various Machine-Generated Text Detection Methods (2025.naacl-srw)
Copied to clipboard
| Challenge: | Recent large-scale emergence of LLMs has left an open space for dealing with consequences, such as plagiarism or the spread of false information on the Internet. |
| Approach: | They evaluate the parsing attack resilience of three machine-generated text detection methods and their ensembles using Random Forest classifiers. |
| Outcome: | The proposed methods yield the strongest results, but suffer significant losses during attacks. |
RAID: A Shared Benchmark for Robust Evaluation of Machine-Generated Text Detectors (2024.acl-long)
Copied to clipboard
Liam Dugan, Alyssa Hwang, Filip Trhlík, Andrew Zhu, Josh Magnus Ludan, Hainiu Xu, Daphne Ippolito, Chris Callison-Burch
| Challenge: | Existing methods for detecting machine-generated text are often insufficiently robust and lack benchmark datasets. |
| Approach: | They evaluate the out-of-domain and adversarial robustness of 8 open- and 4 closed-source detectors using RAID benchmark datasets. |
| Outcome: | The proposed detectors are fooled by adversarial attacks, repetition penalties, and unseen generative models. |
Watermarking PLMs on Classification Tasks by Combining Contrastive Learning with Weight Perturbation (2023.findings-emnlp)
Copied to clipboard
Chenxi Gu, Xiaoqing Zheng, Jianhan Xu, Muling Wu, Cenyuan Zhang, Chengsong Huang, Hua Cai, Xuanjing Huang
| Challenge: | Large pre-trained language models (PLMs) are highly valuable intellectual property due to their expensive training costs. |
| Approach: | They propose to embed backdoors that can be triggered by specific inputs into models by model watermarking. |
| Outcome: | The proposed method can be used to protect the intellectual property of large pre-trained language models without knowledge about downstream tasks. |