论文名称：Visualizing and Understanding Convolutional Networks

论文下载：https://arxiv.org/abs/1311.2901

代码下载：

作者：Matthew D Zeiler, Rob Fergus

发表年份：2013

机构：微软研究院资助的

地位：2013年ImageNet图像分类竞赛冠军模型，提出了一系列可视化卷积神经网络中间层特征的方法，打破神经网络这个黑箱子，并巧妙设置了对照消融实验，从各个角度分析卷积神经网络各层提取的特征及对变换的敏感性

一、 摘要

Large Convolutional Network models have recently demonstrated impressive classifification performance on the ImageNet bench mark (Krizhevsky et al., 2012). However there is no clear understanding of why they perform so well, or how they might be improved. In this paper we address both issues.We introduce a novel visualization technique that gives insight into the function of inter mediate feature layers and the operation of the classififier. Used in a diagnostic role, these visualizations allow us to fifind model architectures that outperform Krizhevsky et al. on the ImageNet classifification benchmark. We also perform an ablation study to discover the performance contribution from different model layers. We show our ImageNet model generalizes well to other datasets: when the softmax classififier is retrained, it convincingly beats the current state-of-the-art results on Caltech-101 and Caltech-256 datasets.

• 在2013年之前，卷积网络的可解释性很差，即很少有人知道卷积网络到底是在做什么，学习什么，卷积网络就是一个黑箱
• 作者提出了一个新颖的可视化技巧，能够可视化卷积网络中间层提取到的特征和最后分类器的行为
• 基于这种技巧，对Alexnet模型进行改进，提出了ZFNet
• 进行了一系列消融（ablation）实验，来探索卷积网络不同层对最终性能的影响

二、网络架构

主要是对AlexNet网络进行了一些修改：

• **在第一个卷积层：**将卷积核从1111改为了77，步伐从4改为了2