TY - GEN

T1 - Learning distance for sequences by learning a ground metric

AU - Su, Bing

AU - Wu, Ying

N1 - Publisher Copyright:
© 2019 International Machine Learning Society (IMLS).

PY - 2019

Y1 - 2019

N2 - Learning distances that operate directly on multidimensional sequences is challenging because such distances are structural by nature and the vectors in sequences arc not independent. Generally, distances for sequences heavily depend on the ground metric between the vectors in sequences. We propose to learn the distance for sequences through learning a ground Mahalanobis metric for the vectors in sequences. The learning samples arc sequences of vectors for which how the ground metric between vectors induces the overall distance is given, and the objective is that the distance induced by the learned ground metric produces large values for sequences from different classes and small values for those from the same class. Wc formulate the metric as a parameter of the distance, bring closer each sequence to an associated virtual sequence w.r.t. the distance to reduce the number of constraints, and develop a general iterative solution for any ground-metric-based sequence distance. Experiments on several sequence datascts demonstrate the effectiveness and efficiency of our method.

AB - Learning distances that operate directly on multidimensional sequences is challenging because such distances are structural by nature and the vectors in sequences arc not independent. Generally, distances for sequences heavily depend on the ground metric between the vectors in sequences. We propose to learn the distance for sequences through learning a ground Mahalanobis metric for the vectors in sequences. The learning samples arc sequences of vectors for which how the ground metric between vectors induces the overall distance is given, and the objective is that the distance induced by the learned ground metric produces large values for sequences from different classes and small values for those from the same class. Wc formulate the metric as a parameter of the distance, bring closer each sequence to an associated virtual sequence w.r.t. the distance to reduce the number of constraints, and develop a general iterative solution for any ground-metric-based sequence distance. Experiments on several sequence datascts demonstrate the effectiveness and efficiency of our method.

UR - http://www.scopus.com/inward/record.url?scp=85078258095&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85078258095&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85078258095

T3 - 36th International Conference on Machine Learning, ICML 2019

SP - 10517

EP - 10534

BT - 36th International Conference on Machine Learning, ICML 2019

PB - International Machine Learning Society (IMLS)

T2 - 36th International Conference on Machine Learning, ICML 2019

Y2 - 9 June 2019 through 15 June 2019

ER -