论文阅读 [TPAMI-2022] Learning Meta-Distance for Sequences by Learning a Ground Metric via Virtual Seque

论文阅读 [TPAMI-2022] Learning Meta-Distance for Sequences by Learning a Ground Metric via Virtual Sequence Regression

论文搜索(studyai.com)

搜索论文: Learning Meta-Distance for Sequences by Learning a Ground Metric via Virtual Sequence Regression

搜索论文: http://www.studyai.com/search/whole-site/?q=Learning+Meta-Distance+for+Sequences+by+Learning+a+Ground+Metric+via+Virtual+Sequence+Regression

关键字(Keywords)

Measurement; Learning systems; Training; Optimization; Neural networks; Machine learning; Kernel; Metric learning; temporal alignment; virtual sequence regression; optimal transport

机器学习

(深度)度量学习

摘要(Abstract)

Distance between sequences is structural by nature because it needs to establish the temporal alignments among the temporally correlated vectors in sequences with varying lengths.

序列之间的距离本质上是结构性的，因为它需要在具有不同长度的序列中的时间相关向量之间建立时间对齐。.

Generally, distances for sequences heavily depend on the ground metric between the vectors in sequences to infer the alignments and hence can be viewed as meta-distances upon the ground metric.

通常，序列的距离在很大程度上依赖于序列中向量之间的地面度量来推断对齐，因此可以被视为地面度量上的元距离。.

Learning such meta-distance from multi-dimensional sequences is appealing but challenging.

从多维序列中学习这种元距离很有吸引力，但也很有挑战性。.

We propose to learn the meta-distance through learning a ground metric for the vectors in sequences.

我们建议通过学习序列中向量的基本度量来学习元距离。.

The learning samples are sequences of vectors for which how the ground metric between vectors induces the meta-distance is given.

学习样本是向量序列，给出了向量之间的地面度量如何诱导元距离。.

The objective is that the meta-distance induced by the learned ground metric produces large values for sequences from different classes and small values for those from the same class.

其目标是，学习的地面度量所产生的元距离为不同类别的序列产生大值，为同一类别的序列产生小值。.

We formulate the ground metric as a parameter of the meta-distance and regress each sequence to an associated pre-generated virtual sequence w.r.t.

我们将地面度量表示为元距离的参数，并将每个序列回归到相关的预生成虚拟序列w.r.t。.

the meta-distance, where the virtual sequences for sequences of different classes are well-separated.

元距离，其中不同类别序列的虚拟序列被很好地分离。.

We develop general iterative solutions to learn both the Mahalanobis metric and the deep metric induced by a neural network for any ground-metric-based sequence distance.

对于任何基于地面度量的序列距离，我们开发了通用迭代解来学习马氏度量和神经网络诱导的深度度量。.

Experiments on several sequence datasets demonstrate the effectiveness and efficiency of the proposed methods…

在多个序列数据集上的实验证明了所提方法的有效性。。.

作者(Authors)

[‘Bing Su’, ‘Ying Wu’]