Deep Validation: Toward Detecting Real-world Corner Cases for Deep Neural Networks

[dsn'19] Deep Validation: Toward Detecting Real-world Corner Cases for Deep Neural Networks

Keywords: DL Robustness, AE detection
Takeaways:

---

Background

1. AE detection

---

Design

1. Motivation

Legitimate input range/probability distribution for every layer is ill-defined, this Is because:

1. the decision functions of these layers are learned on their own rather than manually designed by the developers
2. the classification rules they derive from the training data are encoded in millions of parameters, which are nearly impossible to translate

Key observation: images of different classes can fire different patterns and follow different paths when transferred from one area into another one when going through layers
(相同的label应该有相近的激活路径/隐层表示, 不同的label的也不同)

2. Overview

overview

每类每层train一个OCSVM，然后用signed distance最后算累计（求和）误差，大于一定阈值则判定为corner case

---

Experimental Results

---

Personal Response

+ Strengths：

- Weaknesses:

---

【转载声明】 转载或引用本博客文章请注明出处 -- AISecPaperShare