【译】Attacks against machine learning — an overview
这篇博客文章调查了针对AI（人工智能）系统的攻击技术以及如何防范它们。html 在较高级别，对分类器的攻击能够分为三种类型：node 对抗性输入 ，这是特制的输入，其目的是可靠地错误分类以逃避检测。 对抗性输入包括旨在逃避防病毒的恶意文档，以及试图逃避垃圾邮件过滤器的电子邮件。 数据中毒攻击 ，涉及将训练对抗数据提供给分类器。 咱们观察到的最多见的攻击类型是模型倾斜，攻击者试图以这样的方式污染训练
相关文章
- 1. Practical Black-Box Attacks against Machine Learning
- 2. [paper]Practical Black-Box Attacks against Machine Learning
- 3. Practical Black-Box Attacks against Machine Learning 阅读笔记
- 4. 论文笔记：Membership Inference Attacks Against Machine Learning Models
- 5. 论文解析：Membership Inference Attacks Against Machine Learning Models（一看即懂）
- 6. Learning Machine Learning, Part 1: An Introduction
- 7. An example machine learning notebook
- 8. ÆGIS: Shielding Vulnerable Smart Contracts Against Attacks
- 9. An Introduction to Machine Learning with Python
- 10. Supervised Speech Separation Based on Deep Learning-An Overview
- 更多相关文章...
- • Docker Machine - Docker教程
- • Eclipse 编译项目 - Eclipse 教程
- • Java Agent入门实战（一）-Instrumentation介绍与使用
- • Java Agent入门实战（三）-JVM Attach原理与使用
相关标签/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
欢迎关注本站公众号,获取更多信息
相关文章
- 1. Practical Black-Box Attacks against Machine Learning
- 2. [paper]Practical Black-Box Attacks against Machine Learning
- 3. Practical Black-Box Attacks against Machine Learning 阅读笔记
- 4. 论文笔记：Membership Inference Attacks Against Machine Learning Models
- 5. 论文解析：Membership Inference Attacks Against Machine Learning Models（一看即懂）
- 6. Learning Machine Learning, Part 1: An Introduction
- 7. An example machine learning notebook
- 8. ÆGIS: Shielding Vulnerable Smart Contracts Against Attacks
- 9. An Introduction to Machine Learning with Python
- 10. Supervised Speech Separation Based on Deep Learning-An Overview