On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation.
On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation.
复制标题
DOI：
10.1371/journal.pone.0130140
复制
发表时间：
2015
期刊：
影响因子：
3.7
通讯作者：
Samek W
中科院分区：
文献类型：
--
作者：
Bach S;Binder A;Montavon G;Klauschen F;Müller KR;Samek W
理解和解释自动图像分类系统的分类决策在许多应用中具有很高的价值，因为它允许验证系统的推理并向人类专家提供额外的信息。虽然机器学习方法非常成功地解决了大量的任务，但在大多数情况下，它们的缺点是充当黑盒，不提供任何关于是什么使它们做出特定决定的信息。这项工作提出了一个通用的解决方案，通过像素分解的非线性分类器的理解分类决策的问题。我们介绍了一种方法，允许可视化的贡献，单个像素的预测基于核的分类器的词袋功能和多层神经网络。这些像素贡献可以被可视化为热图，并提供给人类专家，人类专家不仅可以直观地验证分类决策的有效性，还可以进一步分析潜在的感兴趣区域。我们评估了我们的方法，用于在PASCAL VOC 2009图像上训练的分类器，包含几何形状的合成图像数据，MNIST手写数字数据集以及作为Caffe开源包一部分的预训练ImageNet模型。
Understanding and interpreting classification decisions of automated image classification systems is of high value in many applications, as it allows to verify the reasoning of the system and provides additional information to the human expert. Although machine learning methods are solving very successfully a plethora of tasks, they have in most cases the disadvantage of acting as a black box, not providing any information about what made them arrive at a particular decision. This work proposes a general solution to the problem of understanding classification decisions by pixel-wise decomposition of nonlinear classifiers. We introduce a methodology that allows to visualize the contributions of single pixels to predictions for kernel-based classifiers over Bag of Words features and for multilayered neural networks. These pixel contributions can be visualized as heatmaps and are provided to a human expert who can intuitively not only verify the validity of the classification decision, but also focus further analysis on regions of potential interest. We evaluate our method for classifiers trained on PASCAL VOC 2009 images, synthetic image data containing geometric shapes, the MNIST handwritten digits data set and for the pre-trained ImageNet model available as part of the Caffe open source package.
登录
查看更多内容
影响因子：
3.7
作者：
Binder A;Nakajima S;Kloft M;Müller C;Samek W;Brefeld U;Müller KR;Kawanabe M
通讯作者：
Kawanabe M
影响因子：
3.1
作者：
Gevrey, M;Dimopoulos, L;Lek, S
通讯作者：
Lek, S
DOI：
10.1007/978-90-481-3751-0_5
发表时间：
2010-01-01
期刊：
RADAR REMOTE SENSING OF URBAN AREAS
影响因子：
--
作者：
Haensch, Ronny;Hellwich, Olaf
通讯作者：
Hellwich, Olaf
影响因子：
19.5
作者：
Binder, Alexander;Mueller, Klaus-Robert;Kawanabe, Motoaki
通讯作者：
Kawanabe, Motoaki
影响因子：
4.5
作者：
Binder, Alexander;Samek, Wojciech;Kawanabe, Motoaki
通讯作者：
Kawanabe, Motoaki