arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1409.3136cs.LG
收藏

时间序列比对的度量学习

Metric Learning for Temporal Sequence Alignment

Ecole Normale Supérieure(巴黎高等师范学院);Inria Paris Rocquencourt(法国国家信息与自动化研究所巴黎罗克库尔中心)

Damien Garreau, Rémi Lajugie, Sylvain Arlot, Francis Bach

摘要:

在本文中,我们提出了学习马氏距离来执行多变量时间序列的对齐。此任务的学习示例是真实对齐已知的时间序列。我们铸造的对齐问题作为一个结构化的预测任务,并提出现实的损失之间的路线优化是易于处理的。我们提供了在音频到音频的上下文中的真实数据上的实验,在那里我们表明相似性度量的学习导致对齐任务的性能的改善。我们还建议使用这个度量学习框架来执行特征选择,并从基本的音频特征中构建这些特征的组合,以获得更好的对齐性能。

英文摘要:

In this paper, we propose to learn a Mahalanobis distance to perform alignment of multivariate time series. The learning examples for this task are time series for which the true alignment is known. We cast the alignment problem as a structured prediction task, and propose realistic losses between alignments for which the optimization is tractable. We provide experiments on real data in the audio to audio context, where we show that the learning of a similarity measure leads to improvements in the performance of the alignment task. We also propose to use this metric learning framework to perform feature selection and, from basic audio features, build a combination of these with better performance for the alignment.

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。

↑