arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.00032cs.LG
收藏

卷积动态对齐网络用于可解释分类

Convolutional Dynamic Alignment Networks for Interpretable Classifications

MPI for Informatics(马克斯·普朗克信息学研究所);Saarland Informatics Campus(萨尔兰信息学园区);CISPA Helmholtz Center for Information Security(CISPA亥姆霍兹信息安全中心)

Moritz Böhle, Mario Fritz, Bernt Schiele

AI总结:

本文提出卷积动态对齐网络CoDA-Nets,通过动态对齐单元实现输入依赖的线性变换,在保持与ResNet、VGG相当分类性能的同时,提供高质量、可线性分解的内在可解释性。

摘要:

我们提出了一类新的神经网络模型,称为卷积动态对齐网络(CoDA-Nets),它们是具有高度内在可解释性的高性能分类器。其核心构建模块是动态对齐单元(DAUs),这些单元利用与任务相关模式动态对齐的权重向量对其输入进行线性变换。因此,CoDA-Nets通过一系列依赖输入的线性变换来建模分类预测,从而可以将输出线性分解为各个输入贡献。由于DAUs的对齐特性,所得到的贡献图与具有判别力的输入模式对齐。这些模型固有的分解具有很高的视觉质量,并在定量指标下优于现有的归因方法。此外,CoDA-Nets构成了高性能分类器,在CIFAR-10和TinyImagenet等数据集上取得了与ResNet和VGG模型相当的结果。

英文摘要:

We introduce a new family of neural network models called Convolutional Dynamic Alignment Networks (CoDA-Nets), which are performant classifiers with a high degree of inherent interpretability. Their core building blocks are Dynamic Alignment Units (DAUs), which linearly transform their input with weight vectors that dynamically align with task-relevant patterns. As a result, CoDA-Nets model the classification prediction through a series of input-dependent linear transformations, allowing for linear decomposition of the output into individual input contributions. Given the alignment of the DAUs, the resulting contribution maps align with discriminative input patterns. These model-inherent decompositions are of high visual quality and outperform existing attribution methods under quantitative metrics. Further, CoDA-Nets constitute performant classifiers, achieving on par results to ResNet and VGG models on e.g. CIFAR-10 and TinyImagenet.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑