blogs.md

<script type="text/javascript"> function displaySwitch(button,id){ var target=document.getElementById(id); if (target.style.display=="none"){ target.style.display=""; button.innerHTML="HIDE"; } else { target.style.display="none"; button.innerHTML="SHOW"; } } </script>

Blogs

Content

[TOC]

Zhihu Articles

Written for the popularization of science and technologies in china.

• Transformer 的稳健性更好吗？

• Density&Robustness

• 对抗攻击——优化噪声的艺术

• 对抗正则化——盲人摸象还是对症下药？

• 对抗训练——终极数据增强？

• 什么是对抗样本？——稳健性和泛化性的交集

• 对抗样本问题——数据的一小步，模型的一大步

History

Written for my own interests about histories.

• The Ripples of War
• 中国疆域流变

Notes

Comprehensive Notes

Taken while studying some new ideas comprehensively. SHOW

• A Map of Adversarial Example Research
• The Landscape of Adversarial Example
• A Simple Survey for Object Detection
• Orthogonal Convolution
• Normalizing Flow
• Basic Concepts of Support Vector Machine (SVM)

Conceptual Notes

Taken for thoughts. SHOW

• The Categories of Computer Vision
• The Channel in Convolutional Neural Nets

Single Paper Notes

Detailed notaries for important, interesting or useful papers. SHOW

• Adversarial Vulnerability for any Classifier
• PAC Learning with Adversary
• Bag of Tricks for Adversarial Training
• First-order Adversarial Vulnerability
• Shortcut Learning as a Concept
• Are Adversarial Example Inevitable?
• Feature Purification : an Explanation for Adversarial Training
• Higher-level Inductive Biases
• Optimism at the Face of Adversity
• Summarized Survey of Few-Shot Learning
• Self-Attention and Convolution

Dense Paper Notes

Detailed notaries for a set of papers that working on the same problem or idea. SHOW

• Explanation of Robustness and Adversarial Example - 1
• Explanation of Robustness and Adversarial Example - 2
• Interprertation for Robustness and Adversarial Example
• Adversarial Attacks

• Adversarial Example in Object Detection

• Restricted Adversarial Attacks

• Defenses against Adversarial Attacks

  • Adversarial Defense with Ensemble

  • Adversarial Defense at Inference

  • Adversarial Defense by Adversarial Training - 1

  • Adversarial Defense by Adversarial Training - 2

  • Adversarial Defense by Regularization

    • Adversarial Label Smoothing and Logit Squeezing
• Provable Defense
• Robust Structure
• Breach Adversarial Defenses
• Benchmark Adversarial Defenses
• Adversarial Augmentation
• Adversarial and Generative
• Approaches of Few-shot Learning
• Invertible Residual Network
• Interpretation of Neural Networks

Book Notes

Taken for books. SHOW

• How to Build a Computer
• The Evolution of Cryptography

---

BACK TO THE HOMEPAGE