Model-Guardian: Protecting against Data-Free Model Stealing Using Gradient Representations and Deceptive Predictions

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Yang, Yunfei, Chen, Xiaojun, Xuan, Yuexin, Zhao, Zhendong
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915211317870592
author Yang, Yunfei
Chen, Xiaojun
Xuan, Yuexin
Zhao, Zhendong
author_facet Yang, Yunfei
Chen, Xiaojun
Xuan, Yuexin
Zhao, Zhendong
contents Model stealing attack is increasingly threatening the confidentiality of machine learning models deployed in the cloud. Recent studies reveal that adversaries can exploit data synthesis techniques to steal machine learning models even in scenarios devoid of real data, leading to data-free model stealing attacks. Existing defenses against such attacks suffer from limitations, including poor effectiveness, insufficient generalization ability, and low comprehensiveness. In response, this paper introduces a novel defense framework named Model-Guardian. Comprising two components, Data-Free Model Stealing Detector (DFMS-Detector) and Deceptive Predictions (DPreds), Model-Guardian is designed to address the shortcomings of current defenses with the help of the artifact properties of synthetic samples and gradient representations of samples. Extensive experiments on seven prevalent data-free model stealing attacks showcase the effectiveness and superior generalization ability of Model-Guardian, outperforming eleven defense methods and establishing a new state-of-the-art performance. Notably, this work pioneers the utilization of various GANs and diffusion models for generating highly realistic query samples in attacks, with Model-Guardian demonstrating accurate detection capabilities.
format Preprint
id arxiv_https___arxiv_org_abs_2503_18081
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Model-Guardian: Protecting against Data-Free Model Stealing Using Gradient Representations and Deceptive Predictions
Yang, Yunfei
Chen, Xiaojun
Xuan, Yuexin
Zhao, Zhendong
Cryptography and Security
Machine Learning
Model stealing attack is increasingly threatening the confidentiality of machine learning models deployed in the cloud. Recent studies reveal that adversaries can exploit data synthesis techniques to steal machine learning models even in scenarios devoid of real data, leading to data-free model stealing attacks. Existing defenses against such attacks suffer from limitations, including poor effectiveness, insufficient generalization ability, and low comprehensiveness. In response, this paper introduces a novel defense framework named Model-Guardian. Comprising two components, Data-Free Model Stealing Detector (DFMS-Detector) and Deceptive Predictions (DPreds), Model-Guardian is designed to address the shortcomings of current defenses with the help of the artifact properties of synthetic samples and gradient representations of samples. Extensive experiments on seven prevalent data-free model stealing attacks showcase the effectiveness and superior generalization ability of Model-Guardian, outperforming eleven defense methods and establishing a new state-of-the-art performance. Notably, this work pioneers the utilization of various GANs and diffusion models for generating highly realistic query samples in attacks, with Model-Guardian demonstrating accurate detection capabilities.
title Model-Guardian: Protecting against Data-Free Model Stealing Using Gradient Representations and Deceptive Predictions
topic Cryptography and Security
Machine Learning
url https://arxiv.org/abs/2503.18081