Watermark papers

This repo includes papers about the watermarking for text and images.

Text watermark

• Topic-based Watermarks for LLM-Generated Text. Preprint.

  • Alexander Nemecek, Yuzhou Jiang, Erman Ayday

  • https://arxiv.org/abs/2404.02138

• A Statistical Framework of Watermarks for Large Language Models: Pivot, Detection Efficiency and Optimal Rules. Preprint.

  • Xiang Li, Feng Ruan, Huiyuan Wang, Qi Long, Weijie J. Su

  • https://arxiv.org/abs/2404.01245

• WaterJudge: Quality-Detection Trade-off when Watermarking Large Language Models. Preprint.

  • Piotr Molenda, Adian Liusie, Mark J. F. Gales

  • https://arxiv.org/abs/2403.19548

• Duwak: Dual Watermarks in Large Language Models. Preprint.

  • Chaoyi Zhu, Jeroen Galjaard, Pin-Yu Chen, Lydia Y. Chen

  • https://arxiv.org/abs/2403.13000

• Lost in Overlap: Exploring Watermark Collision in LLMs. Preprint.

  • Yiyang Luo, Ke Lin, Chao Gu

  • https://arxiv.org/abs/2403.10020

• WaterMax: breaking the LLM watermark detectability-robustness-quality trade-off. Preprint.

  • Eva Giboulot, Furon Teddy

  • https://arxiv.org/abs/2403.04808

• WARDEN: Multi-Directional Backdoor Watermarks for Embedding-as-a-Service Copyright Protection. Preprint.

  • Anudeex Shetty, Yue Teng, Ke He, Qiongkai Xu

  • https://arxiv.org/abs/2403.01472

• EmMark: Robust Watermarks for IP Protection of Embedded Quantized Large Language Models. Preprint.

  • Ruisi Zhang, Farinaz Koushanfar

  • https://arxiv.org/abs/2402.17938

• Token-Specific Watermarking with Enhanced Detectability and Semantic Coherence for Large Language Models. Preprint.

  • Mingjia Huo, Sai Ashish Somayajula, Youwei Liang, Ruisi Zhang, Farinaz Koushanfar, Pengtao Xie

  • https://arxiv.org/abs/2402.18059

• Attacking LLM Watermarks by Exploiting Their Strengths. Preprint.

  • Qi Pang, Shengyuan Hu, Wenting Zheng, Virginia Smith

  • https://arxiv.org/abs/2402.16187

• Multi-Bit Distortion-Free Watermarking for Large Language Models. preprint.

  • Massieh Kordi Boroujeny, Ya Jiang, Kai Zeng, Brian Mark
  • https://arxiv.org/abs/2402.16578

• Watermarking Makes Language Models Radioactive. Preprint.

  • Tom Sander, Pierre Fernandez, Alain Durmus, Matthijs Douze, Teddy Furon

  • https://arxiv.org/abs/2402.14904

• Can Watermarks Survive Translation? On the Cross-lingual Consistency of Text Watermark for Large Language Models. Preprint.

  • Zhiwei He, Binglin Zhou, Hongkun Hao, Aiwei Liu, Xing Wang, Zhaopeng Tu, Zhuosheng Zhang, Rui Wang

  • https://arxiv.org/abs/2402.14007

• GumbelSoft: Diversified Language Model Watermarking via the GumbelMax-trick. Preprint.

  • Jiayi Fu, Xuandong Zhao, Ruihan Yang, Yuansen Zhang, Jiangjie Chen, Yanghua Xiao

  • https://arxiv.org/abs/2402.12948

• k-SemStamp: A Clustering-Based Semantic Watermark for Detection of Machine-Generated Text. Preprint.

  • Abe Bohan Hou, Jingyu Zhang, Yichen Wang, Daniel Khashabi, Tianxing He

  • https://arxiv.org/abs/2402.11399

• Proving membership in LLM pretraining data via data watermarks. Preprint.

  • Johnny Tian-Zheng Wei, Ryan Yixiang Wang, Robin Jia

  • https://arxiv.org/abs/2402.10892

• Permute-and-Flip: An optimally robust and watermarkable decoder for LLMs. Preprint.

  • Xuandong Zhao, Lei Li, Yu-Xiang Wang
  • https://arxiv.org/abs/2402.05864

• Provably Robust Multi-bit Watermarking for AI-generated Text via Error Correction Code. Preprint.

  • Wenjie Qu, Dong Yin, Zixin He, Wei Zou, Tianyang Tao, Jinyuan Jia, Jiaheng Zhang
  • https://arxiv.org/abs/2401.16820

• Instructional Fingerprinting of Large Language Models. Preprint.

  • Jiashu Xu, Fei Wang, Mingyu Derek Ma, Pang Wei Koh, Chaowei Xiao, Muhao Chen
  • https://arxiv.org/abs/2401.12255

• Adaptive Text Watermark for Large Language Models. Preprint.

  • Yepeng Liu, Yuheng Bu
  • https://arxiv.org/abs/2401.13927

• Excuse me, sir? Your language model is leaking (information) Preprint.

  • Or Zamir

  • https://arxiv.org/abs/2401.10360

• Cross-Attention Watermarking of Large Language Models. ICASSP2024.

  • Folco Bertini Baldassini, Huy H. Nguyen, Ching-Chung Chang, Isao Echizen

  • https://arxiv.org/abs/2401.06829

• Optimizing watermarks for large language models. Preprint.

  • Bram Wouters

  • https://arxiv.org/abs/2312.17295

• Towards Optimal Statistical Watermarking. Preprint.

  • Baihe Huang, Banghua Zhu, Hanlin Zhu, Jason D. Lee, Jiantao Jiao, Michael I. Jordan

  • https://arxiv.org/abs/2312.07930

• A Survey of Text Watermarking in the Era of Large Language Models. Preprint. Survey paper.

  • Aiwei Liu, Leyi Pan, Yijian Lu, Jingjing Li, Xuming Hu, Lijie Wen, Irwin King, Philip S. Yu

  • https://arxiv.org/abs/2312.07913

• On the Learnability of Watermarks for Language Models. Preprint.

  • Chenchen Gu, Xiang Lisa Li, Percy Liang, Tatsunori Hashimoto

  • https://arxiv.org/abs/2312.04469

• New Evaluation Metrics Capture Quality Degradation due to LLM Watermarking. Preprint.

  • Karanpartap Singh, James Zou

  • https://arxiv.org/abs/2312.02382

• Mark My Words: Analyzing and Evaluating Language Model Watermarks. Preprint.

  • Julien Piet, Chawin Sitawarin, Vivian Fang, Norman Mu, David Wagner

  • https://arxiv.org/abs/2312.00273

• I Know You Did Not Write That! A Sampling Based Watermarking Method for Identifying Machine Generated Text. Preprint.

  • Kaan Efe Keleş, Ömer Kaan Gürbüz, Mucahid Kutlu

  • https://arxiv.org/abs/2311.18054

• Improving the Generation Quality of Watermarked Large Language Models via Word Importance Scoring. Preprint

  • Yuhang Li, Yihan Wang, Zhouxing Shi, Cho-Jui Hsieh
  • https://arxiv.org/abs/2311.09668

• Performance Trade-offs of Watermarking Large Language Models. Preprint.

  • Anirudh Ajith, Sameer Singh, Danish Pruthi
  • https://arxiv.org/abs/2311.09816

• X-Mark: Towards Lossless Watermarking Through Lexical Redundancy. Preprint.

  • Liang Chen, Yatao Bian, Yang Deng, Shuaiyi Li, Bingzhe Wu, Peilin Zhao, Kam-fai Wong
  • https://arxiv.org/abs/2311.09832

• WaterBench: Towards Holistic Evaluation of Watermarks for Large Language Models. Preprint.

  • Shangqing Tu, Yuliang Sun, Yushi Bai, Jifan Yu, Lei Hou, Juanzi Li
  • https://arxiv.org/abs/2311.07138
  • Benchmark dataset

• Watermarks in the Sand: Impossibility of Strong Watermarking for Generative Models. Preprint.

  • Hanlin Zhang, Benjamin L. Edelman, Danilo Francati, Daniele Venturi, Giuseppe Ateniese, Boaz Barak

  • https://arxiv.org/abs/2311.04378

• REMARK-LLM: A Robust and Efficient Watermarking Framework for Generative Large Language Models. Preprint.

  • Ruisi Zhang, Shehzeen Samarah Hussain, Paarth Neekhara, Farinaz Koushanfar
  • https://arxiv.org/abs/2310.12362

• Embarrassingly Simple Text Watermarks. Preprint.

  • Ryoma Sato, Yuki Takezawa, Han Bao, Kenta Niwa, Makoto Yamada
  • https://arxiv.org/abs/2310.08920

• Necessary and Sufficient Watermark for Large Language Models. Preprint.

  • Yuki Takezawa, Ryoma Sato, Han Bao, Kenta Niwa, Makoto Yamada
  • https://arxiv.org/abs/2310.00833

• Functional Invariants to Watermark Large Transformers. Preprint.

  • Fernandez Pierre, Couairon Guillaume, Furon Teddy, Douze Matthijs
  • https://arxiv.org/abs/2310.11446

• Watermarking LLMs with Weight Quantization. EMNLP2023 findings.

  • Linyang Li, Botian Jiang, Pengyu Wang, Ke Ren, Hang Yan, Xipeng Qiu
  • https://arxiv.org/abs/2310.11237

• DiPmark: A Stealthy, Efficient and Resilient Watermark for Large Language Models. Preprint.

  • Yihan Wu, Zhengmian Hu, Hongyang Zhang, Heng Huang
  • https://arxiv.org/abs/2310.07710

• A Semantic Invariant Robust Watermark for Large Language Models. Preprint.

  • Aiwei Liu, Leyi Pan, Xuming Hu, Shiao Meng, Lijie Wen
  • https://arxiv.org/abs/2310.06356

• SemStamp: A Semantic Watermark with Paraphrastic Robustness for Text Generation. Preprint.

  • Abe Bohan Hou, Jingyu Zhang, Tianxing He, Yichen Wang, Yung-Sung Chuang, Hongwei Wang, Lingfeng Shen, Benjamin Van Durme, Daniel Khashabi, Yulia Tsvetkov
  • https://arxiv.org/abs/2310.03991

• Advancing Beyond Identification: Multi-bit Watermark for Language Models. Preprint.

  • KiYoon Yoo, Wonhyuk Ahn, Nojun Kwak.
  • https://arxiv.org/abs/2308.00221

• Three Bricks to Consolidate Watermarks for Large Language Models. Preprint.

  • Pierre Fernandez, Antoine Chaffin, Karim Tit, Vivien Chappelier, Teddy Furon.
  • https://arxiv.org/abs/2308.00113

• Towards Codable Text Watermarking for Large Language Models. Preprint.

  • Lean Wang, Wenkai Yang, Deli Chen, Hao Zhou, Yankai Lin, Fandong Meng, Jie Zhou, Xu Sun.
  • https://arxiv.org/abs/2307.15992

• A Private Watermark for Large Language Models. Preprint.

  • Aiwei Liu, Leyi Pan, Xuming Hu, Shu'ang Li, Lijie Wen, Irwin King, Philip S. Yu.
  • https://arxiv.org/abs/2307.16230

• Robust Distortion-free Watermarks for Language Models. Preprint.

  • Rohith Kuditipudi John Thickstun Tatsunori Hashimoto Percy Liang.
  • https://arxiv.org/abs/2307.15593

• Watermarking Conditional Text Generation for AI Detection: Unveiling Challenges and a Semantic-Aware Watermark Remedy. Preprint.

  • Yu Fu, Deyi Xiong, Yue Dong.
  • https://arxiv.org/abs/2307.13808

• Provable Robust Watermarking for AI-Generated Text. Preprint.

  • Xuandong Zhao, Prabhanjan Ananth, Lei Li, Yu-Xiang Wang.
  • https://arxiv.org/abs/2306.17439

• On the Reliability of Watermarks for Large Language Models. Preprint.

  • John Kirchenbauer, Jonas Geiping, Yuxin Wen, Manli Shu, Khalid Saifullah, Kezhi Kong, Kasun Fernando, Aniruddha Saha, Micah Goldblum, Tom Goldstein.
  • https://arxiv.org/abs/2306.04634

• Undetectable Watermarks for Language Models. Preprint.

  • Miranda Christ, Sam Gunn, Or Zamir.
  • https://arxiv.org/abs/2306.09194

• Watermarking Text Data on Large Language Models for Dataset Copyright Protection. Preprint.

  • Yixin Liu, Hongsheng Hu, Xuyun Zhang, Lichao Sun.
  • https://arxiv.org/abs/2305.13257

• Baselines for Identifying Watermarked Large Language Models. Preprint.

  • Leonard Tang, Gavin Uberti, Tom Shlomi.
  • https://arxiv.org/abs/2305.18456

• Who Wrote this Code? Watermarking for Code Generation. Preprint.

  • Taehyun Lee, Seokhee Hong, Jaewoo Ahn, Ilgee Hong, Hwaran Lee, Sangdoo Yun, Jamin Shin, Gunhee Kim.
  • https://arxiv.org/abs/2305.15060

• Robust Multi-bit Natural Language Watermarking through Invariant Features. ACL 2023.

  • KiYoon Yoo, Wonhyuk Ahn, Jiho Jang, Nojun Kwak.
  • https://arxiv.org/abs/2305.01904

• Are You Copying My Model? Protecting the Copyright of Large Language Models for EaaS via Backdoor Watermark. ACL 2023.

  • Wenjun Peng, Jingwei Yi, Fangzhao Wu, Shangxi Wu, Bin Zhu, Lingjuan Lyu, Binxing Jiao, Tong Xu, Guangzhong Sun, Xing Xie.
  • https://arxiv.org/abs/2305.10036

• Watermarking Text Generated by Black-Box Language Models. Preprint.

  • Xi Yang, Kejiang Chen, Weiming Zhang, Chang Liu, Yuang Qi, Jie Zhang, Han Fang, Nenghai Yu.
  • https://arxiv.org/abs/2305.08883

• Protecting Language Generation Models via Invisible Watermarking. ICML 2023.

  • Xuandong Zhao, Yu-Xiang Wang, Lei Li.
  • https://arxiv.org/abs/2302.03162

• A Watermark for Large Language Models. ICML 2023. Outstanding Paper Award

  • John Kirchenbauer, Jonas Geiping, Yuxin Wen, Jonathan Katz, Ian Miers, Tom Goldstein.
  • https://arxiv.org/abs/2301.10226

• Distillation-Resistant Watermarking for Model Protection in NLP. EMNLP 2022

  • Xuandong Zhao, Lei Li, Yu-Xiang Wang.
  • https://arxiv.org/abs/2210.03312

• CATER: Intellectual Property Protection on Text Generation APIs via Conditional Watermarks. NeurIPS 2022

  • Xuanli He, Qiongkai Xu, Yi Zeng, Lingjuan Lyu, Fangzhao Wu, Jiwei Li, Ruoxi Jia.
  • https://arxiv.org/abs/2209.08773

• Adversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hiding. IEEE S&P 2021

  • Sahar Abdelnabi, Mario Fritz.
  • https://arxiv.org/abs/2009.03015

• Watermarking GPT Outputs. slides 2023

  • Scott Aaronson, Hendrik Kirchner
  • https://www.scottaaronson.com/talks/watermark.ppt

• Watermarking the Outputs of Structured Prediction with an Application in Statistical Machine Translation. EMNLP 2011

  • Ashish Venugopal, Jakob Uszkoreit, David Talbot, Franz Och, Juri Ganitkevitch.
  • https://aclanthology.org/D11-1126/

Image watermark

• Flexible and Secure Watermarking for Latent Diffusion Model. MM23.
  • Cheng Xiong, Chuan Qin, Guorui Feng, Xinpeng Zhang
  • https://dl.acm.org/doi/abs/10.1145/3581783.3612448
• Leveraging Optimization for Adaptive Attacks on Image Watermarks. Preprint.
  • Nils Lukas, Abdulrahman Diaa, Lucas Fenaux, Florian Kerschbaum
  • https://arxiv.org/abs/2309.16952
• Catch You Everything Everywhere: Guarding Textual Inversion via Concept Watermarking. Preprint.
  • Weitao Feng, Jiyan He, Jie Zhang, Tianwei Zhang, Wenbo Zhou, Weiming Zhang, Nenghai Yu
  • https://arxiv.org/abs/2309.05940
• Hey That's Mine Imperceptible Watermarks are Preserved in Diffusion Generated Outputs. Preprint.
  • Luke Ditria, Tom Drummond
  • https://arxiv.org/abs/2308.11123
• Generative Watermarking Against Unauthorized Subject-Driven Image Synthesis. Preprint.
  • Yihan Ma, Zhengyu Zhao, Xinlei He, Zheng Li, Michael Backes, Yang Zhang
  • https://arxiv.org/abs/2306.07754
• Invisible Image Watermarks Are Provably Removable Using Generative AI. Preprint.
  • Xuandong Zhao, Kexun Zhang, Zihao Su, Saastha Vasan, Ilya Grishchenko, Christopher Kruegel, Giovanni Vigna, Yu-Xiang Wang, Lei Li.
  • https://arxiv.org/abs/2306.01953
• Tree-Ring Watermarks: Fingerprints for Diffusion Images that are Invisible and Robust. Preprint.
  • Yuxin Wen, John Kirchenbauer, Jonas Geiping, Tom Goldstein.
  • https://arxiv.org/abs/2305.20030
• Evading Watermark based Detection of AI-Generated Content. CCS 2023.
  • Zhengyuan Jiang, Jinghuai Zhang, Neil Zhenqiang Gong.
  • https://arxiv.org/abs/2305.03807
• The Stable Signature: Rooting Watermarks in Latent Diffusion Models. ICCV 2023.
  • Pierre Fernandez, Guillaume Couairon, Hervé Jégou, Matthijs Douze, Teddy Furon.
  • https://arxiv.org/abs/2303.15435
• Watermarking Images in Self-Supervised Latent Spaces. ICASSP 2022.
  • Pierre Fernandez, Alexandre Sablayrolles, Teddy Furon, Hervé Jégou, Matthijs Douze.
  • https://arxiv.org/abs/2112.09581

Contributing to this paper list

First, think about which category the work should belong to.

Second, use the same format as the others to describe the work.