Deep metric learning for sequential data using approximate information

Stefan Thaler; Vlado Menkovski; Milan Petkovic

doi:10.1007/978-3-319-96136-1_22

Deep metric learning for sequential data using approximate information

Stefan Thaler, Vlado Menkovski, Milan Petkovic

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

1 Citaat (Scopus)

2 Downloads (Pure)

Samenvatting

Learning a distance metric provides solutions to many problems where the data exists in a high dimensional space and hand-crafted distance metrics fail to capture its semantical structure. Methods based on deep neural networks such as Siamese or Triplet networks have been developed for learning such metrics. In this paper we present a metric learning method for sequence data based on a RNN-based triplet network. We posit that this model can be trained efficiently with regards to labels by using Jaccard distance as a proxy distance metric. We empirically demonstrate the performance and efficiency of the approach on three different computer log-line datasets.

Originele taal-2	Engels
Titel	Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings
Redacteuren	Petra Perner
Plaats van productie	Cham
Uitgeverij	Springer
Pagina's	269-282
Aantal pagina's	14
ISBN van elektronische versie	978-3-319-96136-1
ISBN van geprinte versie	978-3-319-96135-4
DOI's	https://doi.org/10.1007/978-3-319-96136-1_22
Status	Gepubliceerd - 8 jul. 2018
Evenement	14th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2018 - New York, Verenigde Staten van Amerika Duur: 15 jul. 2018 → 19 jul. 2018

Publicatie series

Naam	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10934 LNAI
ISSN van geprinte versie	0302-9743
ISSN van elektronische versie	1611-3349

Congres

Congres	14th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2018
Land/Regio	Verenigde Staten van Amerika
Stad	New York
Periode	15/07/18 → 19/07/18

Financiering

Acknowledgment. The work presented in this paper is part of a project which has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 780495.

Toegang tot document

10.1007/978-3-319-96136-1_22

Andere bestanden en links

Link to publication in Scopus

Citeer dit

Thaler, S., Menkovski, V., & Petkovic, M. (2018). Deep metric learning for sequential data using approximate information. In P. Perner (editor), Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings (blz. 269-282). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10934 LNAI). Springer. https://doi.org/10.1007/978-3-319-96136-1_22

Thaler, Stefan ; Menkovski, Vlado ; Petkovic, Milan. / Deep metric learning for sequential data using approximate information. Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings. redacteur / Petra Perner. Cham : Springer, 2018. blz. 269-282 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{1a812ec0bf7f429f9ccd5650a92bff64,

title = "Deep metric learning for sequential data using approximate information",

abstract = "Learning a distance metric provides solutions to many problems where the data exists in a high dimensional space and hand-crafted distance metrics fail to capture its semantical structure. Methods based on deep neural networks such as Siamese or Triplet networks have been developed for learning such metrics. In this paper we present a metric learning method for sequence data based on a RNN-based triplet network. We posit that this model can be trained efficiently with regards to labels by using Jaccard distance as a proxy distance metric. We empirically demonstrate the performance and efficiency of the approach on three different computer log-line datasets.",

keywords = "Deep learning, Efficient metric learning, Triplet network",

author = "Stefan Thaler and Vlado Menkovski and Milan Petkovic",

year = "2018",

month = jul,

day = "8",

doi = "10.1007/978-3-319-96136-1_22",

language = "English",

isbn = "978-3-319-96135-4",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "269--282",

editor = "Petra Perner",

booktitle = "Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings",

address = "Germany",

note = "14th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2018 ; Conference date: 15-07-2018 Through 19-07-2018",

}

Thaler, S, Menkovski, V & Petkovic, M 2018, Deep metric learning for sequential data using approximate information. in P Perner (uitgave), Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10934 LNAI, Springer, Cham, blz. 269-282, 14th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2018, New York, Verenigde Staten van Amerika, 15/07/18. https://doi.org/10.1007/978-3-319-96136-1_22

Deep metric learning for sequential data using approximate information. / Thaler, Stefan; Menkovski, Vlado ; Petkovic, Milan.
Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings. uitgave / Petra Perner. Cham: Springer, 2018. blz. 269-282 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10934 LNAI).

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

TY - GEN

T1 - Deep metric learning for sequential data using approximate information

AU - Thaler, Stefan

AU - Menkovski, Vlado

AU - Petkovic, Milan

PY - 2018/7/8

Y1 - 2018/7/8

N2 - Learning a distance metric provides solutions to many problems where the data exists in a high dimensional space and hand-crafted distance metrics fail to capture its semantical structure. Methods based on deep neural networks such as Siamese or Triplet networks have been developed for learning such metrics. In this paper we present a metric learning method for sequence data based on a RNN-based triplet network. We posit that this model can be trained efficiently with regards to labels by using Jaccard distance as a proxy distance metric. We empirically demonstrate the performance and efficiency of the approach on three different computer log-line datasets.

AB - Learning a distance metric provides solutions to many problems where the data exists in a high dimensional space and hand-crafted distance metrics fail to capture its semantical structure. Methods based on deep neural networks such as Siamese or Triplet networks have been developed for learning such metrics. In this paper we present a metric learning method for sequence data based on a RNN-based triplet network. We posit that this model can be trained efficiently with regards to labels by using Jaccard distance as a proxy distance metric. We empirically demonstrate the performance and efficiency of the approach on three different computer log-line datasets.

KW - Deep learning

KW - Efficient metric learning

KW - Triplet network

UR - http://www.scopus.com/inward/record.url?scp=85050538775&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-96136-1_22

DO - 10.1007/978-3-319-96136-1_22

M3 - Conference contribution

AN - SCOPUS:85050538775

SN - 978-3-319-96135-4

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 269

EP - 282

BT - Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings

A2 - Perner, Petra

PB - Springer

CY - Cham

T2 - 14th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2018

Y2 - 15 July 2018 through 19 July 2018

ER -

Thaler S, Menkovski V , Petkovic M. Deep metric learning for sequential data using approximate information. In Perner P, redacteur, Machine Learning and Data Mining in Pattern Recognition - 14th International Conference, MLDM 2018, Proceedings. Cham: Springer. 2018. blz. 269-282. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-96136-1_22

Deep metric learning for sequential data using approximate information

Samenvatting

Publicatie series

Congres

Financiering

Toegang tot document

Andere bestanden en links

Vingerafdruk

Citeer dit