paper-with-me

홈 › Papers

Theoretically Principled Trade-off for Stateful Defenses against Query-Based Black-Box Attacks

2023-07-30 · Ashish Hooda, Neal Mangaokar, Ryan Feng, Kassem Fawaz, Somesh Jha, Atul Prakash

Adversarial examples threaten the integrity of machine learning systems with alarming success rates even under constrained black-box conditions. Stateful defenses have emerged as an effective countermeasure, detecting potential attacks by maintaining a buffer of recent queries and detecting new queries that are too similar. However, these defenses fundamentally pose a trade-off between attack detection and false positive rates, and this trade-off is typically optimized by hand-picking feature extractors and similarity thresholds that empirically work well. There is little current understanding as to the formal limits of this trade-off and the exact properties of the feature extractors/underlying problem domain that influence it. This work aims to address this gap by offering a theoretical characterization of the trade-off between detection and false positive rates for stateful defenses. We provide upper bounds for detection rates of a general class of feature extractors and analyze the impact of this trade-off on the convergence of black-box attacks. We then support our theoretical findings with empirical evaluations across multiple datasets and stateful defenses.

📄 PDF Abstract BibTeX arXiv:2307.16331

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

MalProtect: Stateful Defense Against Adversarial Query Attacks in ML-based Malware Detection

2023-02-21 · Aqib Rashid, Jose Such

ML models are known to be vulnerable to adversarial query attacks. In these attacks, queries are iteratively perturbed towards a particular class without any knowledge of the target model besides its output. The prevalen…

Malware DetectionOut-of-Distribution Detection

Stateful Defenses for Machine Learning Models Are Not Yet Secure Against Black-box Attacks

2023-03-11 · Ryan Feng, Ashish Hooda, Neal Mangaokar, Kassem Fawaz 외

Recent work has proposed stateful defense models (SDMs) as a compelling strategy to defend against a black-box attacker who only has query access to the model, as is common for online machine learning platforms. Such sta…

Benchmarking Misuse Mitigation Against Covert Adversaries

2025-06-06 · Davis Brown, Mahdi Sabbaghi, Luze Sun, Alexander Robey 외

Existing language model safety evaluations focus on overt attacks and low-stakes tasks. Realistic attackers can subvert current safeguards by requesting help on small, benign-seeming tasks across many independent queries…

BenchmarkingLanguage ModelingLanguage Modelling

Theoretically Principled Trade-off between Robustness and Accuracy

2019-01-24 · Hongyang Zhang, Yaodong Yu, Jiantao Jiao, Eric P. Xing 외

We identify a trade-off between robustness and accuracy that serves as a guiding principle in the design of defenses against adversarial examples. Although this problem has been widely studied empirically, much remains u…

Adversarial AttackAdversarial DefenseAdversarial RobustnessGeneral Classification

Decomposition Attacks Across Unlinkable Identities: Limits of Stateful Defenses for LLM Services

2026-08-18 · Bowen Sun, Zhengyue Zhao, Xiaogeng Liu, Yinzhi Cao 외 arxiv

Most large language model services use stateless defenses, which judge only the current request, to refuse harmful tasks. Decomposition attacks exploit this limitation by splitting a harmful task into individually permis…