paperbot · PL 论文追踪

全部论文 待读 收藏

P2 DeepSearch: a simple and effective blackbox attack for deep neural networks

DeepSearch:一种简单有效的深度神经网络黑盒攻击
FSE Proceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering2020匹配度 1/10 · 该论文研究的是针对图像分类器的黑盒对抗攻击(基于 fuzzing 生成对抗样本),属于机器学习安全/测试方向,与用户的 PL↔AI 静态分析、Agentic 程序分析、浮点/贝叶斯程序分析及其与 LLM 的交叉均无实质关联;仅标签中笼统的“测试、调试与修复”表面相近,但方法与问题均不契合。引用 33 inbox
Fuyuan Zhang, Sankalan Pal Chowdhury, Maria Christakis
crossref: 2026-09-12

AI 速览

一句话:提出基于 fuzzing 的查询高效黑盒攻击 DeepSearch,能更有效找到图像分类器的对抗输入。

深度神经网络在图像分类任务上表现优异,但容易受到对抗攻击;现有生成对抗输入的方法包括白盒与黑盒攻击。本文提出 DeepSearch,一种基于 fuzzing 的新型黑盒攻击,专用于图像分类器,并强调查询效率。尽管方法简单,实验表明 DeepSearch 在寻找对抗输入方面比当前先进的黑盒方法更有效,同时能够生成相比这些方法更隐蔽的对抗输入。该工作为黑盒对抗攻击提供了一种简洁且高效的方案。

问题
如何以黑盒、查询高效的方式为图像分类器生成对抗输入,并提升攻击有效性与扰动隐蔽性。
方法
提出 DeepSearch,将 fuzzing 用于黑盒对抗样本搜索,通过查询目标模型生成对抗输入。
评估
在图像分类任务上与当前先进黑盒攻击方法比较,显示其在找到对抗输入和生成更隐蔽对抗输入方面更优。
原文摘要(Abstract)

Although deep neural networks have been very successful in image-classification tasks, they are prone to adversarial attacks. To generate adversarial inputs, there has emerged a wide variety of techniques, such as black- and whitebox attacks for neural networks. In this paper, we present DeepSearch, a novel fuzzing-based, query-efficient, blackbox attack for image classifiers. Despite its simplicity, DeepSearch is shown to be more effective in finding adversarial inputs than state-of-the-art blackbox approaches. DeepSearch is additionally able to generate the most subtle adversarial inputs in comparison to these approaches.

链接与引用

DOI 原文 · PDF(开放获取) ·

BibTeX
@inproceedings{paperbot5800,
  title = {DeepSearch: a simple and effective blackbox attack for deep neural networks},
  author = {Fuyuan Zhang and Sankalan Pal Chowdhury and Maria Christakis},
  booktitle = {FSE},
  year = {2020},
  doi = {10.1145/3368089.3409750}
}