DTWNet: a Dynamic Time Warping Network
DTWNet: a Dynamic Time Warping Network
复制标题
DOI：
--
复制
发表时间：
2019
影响因子：
14.9
通讯作者：
Xingyu Cai;Tingyang Xu;Jinfeng Yi;Junzhou Huang;S. Rajasekaran
中科院分区：
文献类型：
--
作者：
Xingyu Cai;Tingyang Xu;Jinfeng Yi;Junzhou Huang;S. Rajasekaran
动态时间规整（DTW）作为一种相似性度量方法在各个领域得到了广泛的应用。由于其对时间轴上的翘曲的不变性，DTW提供了更有意义的两个信号之间的差异测量比其他距离测量。在本文中，我们提出了一个新的人工神经网络组件。与以前成功使用DTW作为损失函数相比，所提出的框架利用DTW来获得更好的特征提取。首次从理论上分析了DTW的损失，并提出了一种随机反向传播的方法来提高DTW学习的精度和效率。我们还表明，所提出的框架可以作为一个数据分析工具来执行数据分解。
Dynamic Time Warping (DTW) is widely used as a similarity measure in various domains. Due to its invariance against warping in the time axis, DTW provides more meaningful discrepancy measurements between two signals than other dis- tance measures. In this paper, we propose a novel component in an artificial neural network. In contrast to the previous successful usage of DTW as a loss function, the proposed framework leverages DTW to obtain a better feature extraction. For the first time, the DTW loss is theoretically analyzed, and a stochastic backpropogation scheme is proposed to improve the accuracy and efficiency of the DTW learning. We also demonstrate that the proposed framework can be used as a data analysis tool to perform data decomposition.