paper-with-me

홈 › Papers

Defense Against Model Stealing Based on Account-Aware Distribution Discrepancy

2025-03-16 · Jian-Ping Mei, Weibin Zhang, Jie Chen, Xuyun Zhang, Tiantian Zhu

Malicious users attempt to replicate commercial models functionally at low cost by training a clone model with query responses. It is challenging to timely prevent such model-stealing attacks to achieve strong protection and maintain utility. In this paper, we propose a novel non-parametric detector called Account-aware Distribution Discrepancy (ADD) to recognize queries from malicious users by leveraging account-wise local dependency. We formulate each class as a Multivariate Normal distribution (MVN) in the feature space and measure the malicious score as the sum of weighted class-wise distribution discrepancy. The ADD detector is combined with random-based prediction poisoning to yield a plug-and-play defense module named D-ADD for image classification models. Results of extensive experimental studies show that D-ADD achieves strong defense against different types of attacks with little interference in serving benign users for both soft and hard-label settings.

📄 PDF Abstract BibTeX arXiv:2503.12497

Code (1)

AI-EXP-group/D-ADD 공식 구현 pytorch

Tasks

image-classificationImage Classification

Similar Papers 제목 키워드 기반

Prediction Poisoning: Towards Defenses Against DNN Model Stealing Attacks

2019-06-26 · ICLR 2020 1 · Tribhuvanesh Orekondy, Bernt Schiele, Mario Fritz

High-performance Deep Neural Networks (DNNs) are increasingly deployed in many real-world applications e.g., cloud prediction APIs. Recent advances in model functionality stealing attacks via black-box access (i.e., inpu…

Autonomous VehiclesPrediction

Bucks for Buckets (B4B): Active Defenses Against Stealing Encoders

2023-10-12 · NeurIPS 2023 11

Machine Learning as a Service (MLaaS) APIs provide ready-to-use and high-utility encoders that generate vector representations for given inputs. Since these encoders are very costly to train, they become lucrative target…

Adversarial Sparse Teacher: Defense Against Distillation-Based Model Stealing Attacks Using Adversarial Examples

2024-03-08 · Eda Yilmaz, Hacer Yalim Keles

We introduce Adversarial Sparse Teacher (AST), a robust defense method against distillation-based model stealing attacks. Our approach trains a teacher model using adversarial examples to produce sparse logit responses a…

Knowledge Distillation

Defending Against Model Stealing Attacks with Adaptive Misinformation

2019-11-16 · CVPR 2020 6 · Sanjay Kariyappa, Moinuddin K. Qureshi

Deep Neural Networks (DNNs) are susceptible to model stealing attacks, which allows a data-limited adversary with no knowledge of the training dataset to clone the functionality of a target model, just by using black-box…

Misinformationmodel

Efficient Defense Against Model Stealing Attacks on Convolutional Neural Networks

2023-09-04 · Kacem Khaled, Mouna Dhaouadi, Felipe Gohring de Magalhães, Gabriela Nicolescu

Model stealing attacks have become a serious concern for deep learning models, where an attacker can steal a trained model by querying its black-box API. This can lead to intellectual property theft and other security an…