mlsanityData to 7 Oct 2026

Neural Honeytrace: Plug&Play Watermarking Framework against Model Extraction Attacks

arXiv 2501.09328 ↗First posted 16 Jan 2025cs.CR cs.AIcited by 0

On the map: LLM safety and security → Watermarking generated content and models

Abstract (arXiv)

Triggerable watermarking enables model owners to assert ownership against model extraction attacks. However, most existing approaches require additional training, which limits post-deployment flexibility, and the lack of clear theoretical foundations makes them vulnerable to adaptive attacks. In this paper, we propose Neural Honeytrace, a plug-and-play watermarking framework that operates without retraining. We redefine the watermark transmission mechanism from an information perspective, designing a training-free multi-step transmission strategy that leverages the long-tailed effect of backdoor learning to achieve efficient and robust watermark embedding. Extensive experiments demonstrate that Neural Honeytrace reduces the average number of queries required for a worst-case t-test-based ownership verification to as low as 2% of existing methods, while incurring zero training cost.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Watermarking generated content and models, one of 12 directions in LLM safety and security. Builds on 26 of its 41 references that are in the map. First posted on arXiv 16 Jan 2025.

Builds on

26 of 41 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024MEA-Defender: A Robust Watermark against Model Extraction Attack26 Jan

Other references · 25

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  4. 2015Deep Learning and the Information Bottleneck Principlecited by 793
  5. 2016Stealing Machine Learning Models via Prediction APIscited by 602
  6. 2018Knockoff Nets: Stealing Functionality of Black-Box Modelscited by 285
  7. 2019High Accuracy and High Fidelity Extraction of Neural Networkscited by 208
  8. 2018PRADA: Protecting against DNN Model Stealing Attackscited by 159
  9. 2020Entangled Watermarks as a Defense against Model Extractioncited by 144
  10. 2020Data-Free Model Extractioncited by 121
  11. 2020MAZE: Data-Free Model Stealing Attack Using Zeroth-Order Gradient Estimationcited by 99
  12. 2019DAWN: Dynamic Adversarial Watermarking of Neural Networkscited by 94
13 more
  1. 2022Membership Inference Attacks by Exploiting Loss Trajectorycited by 93
  2. 2018Exploring Connections Between Active Learning and Model Extractioncited by 84
  3. 2019Prediction Poisoning: Towards Defenses Against DNN Model Stealing Attackscited by 79
  4. 2022Data-free Backdoor Removal based on Channel Lipschitznesscited by 77
  5. 2021SoK: How Robust is Image Classification Deep Neural Network Watermarking? (Extended Version)cited by 67
  6. 2022Towards Data-Free Model Stealing in a Hard Label Settingcited by 57
  7. 2019Defending Against Model Stealing Attacks with Adaptive Misinformationcited by 55
  8. 2019Robust Watermarking of Neural Network with Exponential Weightingcited by 48
  9. 2022SSLGuard: A Watermarking Scheme for Self-supervised Learning Pre-trained Encoderscited by 43
  10. 2017Model Extraction Warning in MLaaS Paradigmcited by 39
  11. 2021MEGEX: Data-Free Model Extraction Attack against Gradient-Based Explainable AIcited by 35
  12. 2022How to Steer Your Adversary: Targeted and Efficient Model Stealing Defenses with Gradient Redirectioncited by 21
  13. 2022SSL-WM: A Black-Box Watermarking Approach for Encoders Pre-trained by Self-supervised Learningcited by 9

Built on by

Not cited yet.