Proving membership in LLM pretraining data via data watermarks

• URL: http://arxiv.org/abs/2402.10892v3
• Date: Sat, 17 Aug 2024 10:04:53 GMT
• Title: Proving membership in LLM pretraining data via data watermarks
• Authors: Johnny Tian-Zheng Wei, Ryan Yixiang Wang, Robin Jia,
• Abstract summary: This work proposes using data watermarks to enable principled detection with only black-box model access. We study two watermarks: one that inserts random sequences, and another that randomly substitutes characters with Unicode lookalikes. We show that we can robustly detect hashes from BLOOM-176B's training data, as long as they occurred at least 90 times.
• Score: 20.57538940552033
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Detecting whether copyright holders' works were used in LLM pretraining is poised to be an important problem. This work proposes using data watermarks to enable principled detection with only black-box model access, provided that the rightholder contributed multiple training documents and watermarked them before public release. By applying a randomly sampled data watermark, detection can be framed as hypothesis testing, which provides guarantees on the false detection rate. We study two watermarks: one that inserts random sequences, and another that randomly substitutes characters with Unicode lookalikes. We first show how three aspects of watermark design -- watermark length, number of duplications, and interference -- affect the power of the hypothesis test. Next, we study how a watermark's detection strength changes under model and dataset scaling: while increasing the dataset size decreases the strength of the watermark, watermarks remain strong if the model size also increases. Finally, we view SHA hashes as natural watermarks and show that we can robustly detect hashes from BLOOM-176B's training data, as long as they occurred at least 90 times. Together, our results point towards a promising future for data watermarks in real world use.

Related papers

• TabularMark: Watermarking Tabular Datasets for Machine Learning [20.978995194849297]
  We propose a hypothesis testing-based watermarking scheme, TabularMark. Data noise partitioning is utilized for data perturbation during embedding. Experiments on real-world and synthetic datasets demonstrate the superiority of TabularMark in detectability, non-intrusiveness, and robustness.
  arXiv  Detail & Related papers  (2024-06-21T02:58:45Z)
• Black-Box Detection of Language Model Watermarks [1.9374282535132377]
  We develop rigorous statistical tests to detect the presence of all three most popular watermarking scheme families using only a limited number of black-box queries. Our findings indicate that current watermarking schemes are more detectable than previously believed, and that obscuring the fact that a watermark was deployed may not be a viable way for providers to protect against adversaries.
  arXiv  Detail & Related papers  (2024-05-28T08:41:30Z)
• Unbiased Watermark for Large Language Models [67.43415395591221]
  This study examines how significantly watermarks impact the quality of model-generated outputs. It is possible to integrate watermarks without affecting the output probability distribution. The presence of watermarks does not compromise the performance of the model in downstream tasks.
  arXiv  Detail & Related papers  (2023-09-22T12:46:38Z)
• Towards Robust Model Watermark via Reducing Parametric Vulnerability [57.66709830576457]
  backdoor-based ownership verification becomes popular recently, in which the model owner can watermark the model. We propose a mini-max formulation to find these watermark-removed models and recover their watermark behavior. Our method improves the robustness of the model watermarking against parametric changes and numerous watermark-removal attacks.
  arXiv  Detail & Related papers  (2023-09-09T12:46:08Z)
• An Unforgeable Publicly Verifiable Watermark for Large Language Models [84.2805275589553]
  Current watermark detection algorithms require the secret key used in the watermark generation process, making them susceptible to security breaches and counterfeiting during public detection. We propose an unforgeable publicly verifiable watermark algorithm named UPV that uses two different neural networks for watermark generation and detection, instead of using the same key at both stages.
  arXiv  Detail & Related papers  (2023-07-30T13:43:27Z)
• Did You Train on My Dataset? Towards Public Dataset Protection with Clean-Label Backdoor Watermarking [54.40184736491652]
  We propose a backdoor-based watermarking approach that serves as a general framework for safeguarding public-available data. By inserting a small number of watermarking samples into the dataset, our approach enables the learning model to implicitly learn a secret function set by defenders. This hidden function can then be used as a watermark to track down third-party models that use the dataset illegally.
  arXiv  Detail & Related papers  (2023-03-20T21:54:30Z)
• On Function-Coupled Watermarks for Deep Neural Networks [15.478746926391146]
  We propose a novel DNN watermarking solution that can effectively defend against watermark removal attacks. Our key insight is to enhance the coupling of the watermark and model functionalities. Results show a 100% watermark authentication success rate under aggressive watermark removal attacks.
  arXiv  Detail & Related papers  (2023-02-08T05:55:16Z)
• A Watermark for Large Language Models [84.95327142027183]
  We propose a watermarking framework for proprietary language models. The watermark can be embedded with negligible impact on text quality. It can be detected using an efficient open-source algorithm without access to the language model API or parameters.
  arXiv  Detail & Related papers  (2023-01-24T18:52:59Z)
• Open-sourced Dataset Protection via Backdoor Watermarking [87.15630326131901]
  We propose a emphbackdoor embedding based dataset watermarking method to protect an open-sourced image-classification dataset. We use a hypothesis test guided method for dataset verification based on the posterior probability generated by the suspicious third-party model.
  arXiv  Detail & Related papers  (2020-10-12T16:16:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.