Basis vector orthogonalization for an improved kernel gradient matching pursuit method

Yotaro Kubo, Shinji Watanabe, Atsushi Nakamura, Simon Wiesler, Ralf Schlueter, Hermann Ney

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

With the aim of achieving a computationally efficient optimization of kernel-based probabilistic models for various problems, such as sequential pattern recognition, we have already developed the kernel gradient matching pursuit method as an approximation technique for kernel-based classification. The conventional kernel gradient matching pursuit method approximates the optimal parameter vector by using a linear combination of a small number of basis vectors. In this paper, we propose an improved kernel gradient matching pursuit method that introduces orthogonality constraints to the obtained basis vector set. We verified the efficiency of the proposed method by conducting recognition experiments based on handwritten image datasets and speech datasets. We realized a scalable kernel optimization that incorporated various models, handled very high-dimensional features (>100 K features), and enabled the use of large scale datasets (> 10 M samples).

Original languageEnglish
Title of host publication2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings
Pages1909-1912
Number of pages4
DOIs
Publication statusPublished - 2012
Externally publishedYes
Event2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Kyoto
Duration: 2012 Mar 252012 Mar 30

Other

Other2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012
CityKyoto
Period12/3/2512/3/30

Fingerprint

Pattern recognition
Experiments
Statistical Models

Keywords

  • hidden Markov models
  • Kernel methods
  • orthogonal expansion
  • speech recognition

ASJC Scopus subject areas

  • Signal Processing
  • Software
  • Electrical and Electronic Engineering

Cite this

Kubo, Y., Watanabe, S., Nakamura, A., Wiesler, S., Schlueter, R., & Ney, H. (2012). Basis vector orthogonalization for an improved kernel gradient matching pursuit method. In 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings (pp. 1909-1912). [6288277] https://doi.org/10.1109/ICASSP.2012.6288277

Basis vector orthogonalization for an improved kernel gradient matching pursuit method. / Kubo, Yotaro; Watanabe, Shinji; Nakamura, Atsushi; Wiesler, Simon; Schlueter, Ralf; Ney, Hermann.

2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings. 2012. p. 1909-1912 6288277.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kubo, Y, Watanabe, S, Nakamura, A, Wiesler, S, Schlueter, R & Ney, H 2012, Basis vector orthogonalization for an improved kernel gradient matching pursuit method. in 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings., 6288277, pp. 1909-1912, 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012, Kyoto, 12/3/25. https://doi.org/10.1109/ICASSP.2012.6288277
Kubo Y, Watanabe S, Nakamura A, Wiesler S, Schlueter R, Ney H. Basis vector orthogonalization for an improved kernel gradient matching pursuit method. In 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings. 2012. p. 1909-1912. 6288277 https://doi.org/10.1109/ICASSP.2012.6288277
Kubo, Yotaro ; Watanabe, Shinji ; Nakamura, Atsushi ; Wiesler, Simon ; Schlueter, Ralf ; Ney, Hermann. / Basis vector orthogonalization for an improved kernel gradient matching pursuit method. 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings. 2012. pp. 1909-1912
@inproceedings{6cf3cf1c7ead40e599146e6bce268fc0,
title = "Basis vector orthogonalization for an improved kernel gradient matching pursuit method",
abstract = "With the aim of achieving a computationally efficient optimization of kernel-based probabilistic models for various problems, such as sequential pattern recognition, we have already developed the kernel gradient matching pursuit method as an approximation technique for kernel-based classification. The conventional kernel gradient matching pursuit method approximates the optimal parameter vector by using a linear combination of a small number of basis vectors. In this paper, we propose an improved kernel gradient matching pursuit method that introduces orthogonality constraints to the obtained basis vector set. We verified the efficiency of the proposed method by conducting recognition experiments based on handwritten image datasets and speech datasets. We realized a scalable kernel optimization that incorporated various models, handled very high-dimensional features (>100 K features), and enabled the use of large scale datasets (> 10 M samples).",
keywords = "hidden Markov models, Kernel methods, orthogonal expansion, speech recognition",
author = "Yotaro Kubo and Shinji Watanabe and Atsushi Nakamura and Simon Wiesler and Ralf Schlueter and Hermann Ney",
year = "2012",
doi = "10.1109/ICASSP.2012.6288277",
language = "English",
isbn = "9781467300469",
pages = "1909--1912",
booktitle = "2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings",

}

TY - GEN

T1 - Basis vector orthogonalization for an improved kernel gradient matching pursuit method

AU - Kubo, Yotaro

AU - Watanabe, Shinji

AU - Nakamura, Atsushi

AU - Wiesler, Simon

AU - Schlueter, Ralf

AU - Ney, Hermann

PY - 2012

Y1 - 2012

N2 - With the aim of achieving a computationally efficient optimization of kernel-based probabilistic models for various problems, such as sequential pattern recognition, we have already developed the kernel gradient matching pursuit method as an approximation technique for kernel-based classification. The conventional kernel gradient matching pursuit method approximates the optimal parameter vector by using a linear combination of a small number of basis vectors. In this paper, we propose an improved kernel gradient matching pursuit method that introduces orthogonality constraints to the obtained basis vector set. We verified the efficiency of the proposed method by conducting recognition experiments based on handwritten image datasets and speech datasets. We realized a scalable kernel optimization that incorporated various models, handled very high-dimensional features (>100 K features), and enabled the use of large scale datasets (> 10 M samples).

AB - With the aim of achieving a computationally efficient optimization of kernel-based probabilistic models for various problems, such as sequential pattern recognition, we have already developed the kernel gradient matching pursuit method as an approximation technique for kernel-based classification. The conventional kernel gradient matching pursuit method approximates the optimal parameter vector by using a linear combination of a small number of basis vectors. In this paper, we propose an improved kernel gradient matching pursuit method that introduces orthogonality constraints to the obtained basis vector set. We verified the efficiency of the proposed method by conducting recognition experiments based on handwritten image datasets and speech datasets. We realized a scalable kernel optimization that incorporated various models, handled very high-dimensional features (>100 K features), and enabled the use of large scale datasets (> 10 M samples).

KW - hidden Markov models

KW - Kernel methods

KW - orthogonal expansion

KW - speech recognition

UR - http://www.scopus.com/inward/record.url?scp=84867619247&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867619247&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2012.6288277

DO - 10.1109/ICASSP.2012.6288277

M3 - Conference contribution

SN - 9781467300469

SP - 1909

EP - 1912

BT - 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Proceedings

ER -