几种深度学习可视化方法（针对层级化特征表示）的辨析

几种深度学习层级化特征学习可视化图的辨析

Hierarchical Representations
Visualization

深度学习（deep learning，DL）的强大得益于其层级化的特征学习能力。常有以下几种图“可视化”DL不同层学习到的特征，以说明模型是如何逐级抽象/提取特征的。
然而一直困惑的是，这几类图有细微差别，甚至矛盾。之前也就人云亦云，今日细究之。

它们分别是

1. 卷积核/滤波器的可视化（无监督卷积DBN）
2. 反卷积
3. 激活最大化（输入空间）
4. 特征图

Lee H, Grosse R, Ranganath R, et al. Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations[C]//Proceedings of the 26th annual international conference on machine learning. 2009: 609-616.

Zeiler M D, Fergus R. Visualizing and understanding convolutional networks[C]//European conference on computer vision. Springer, Cham, 2014: 818-833.

Donglai Wei, Bolei Zhou, Antonio Torralba, William T. Freeman（2015）: mNeuron: A Matlab Plugin to Visualize Neurons from Deep Models.
https://donglaiw.github.io/proj/mneuron/index.html
https://github.com/donglaiw/mNeuron
https://distill.pub/2017/feature-visualization/

激活最大化其他论文：
Simonyan K, Vedaldi A, Zisserman A. Deep inside convolutional networks: Visualising image classification models and saliency maps[C]//In Workshop at International Conference on Learning Representations. 2014.
Erhan D, Bengio Y, Courville A, et al. Visualizing higher-layer features of a deep network[J]. University of Montreal, 2009, 1341(3): 1.

Taigman Y, Yang M, Ranzato M A, et al. Deepface: Closing the gap to human-level performance in face verification[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2014: 1701-1708.