Currently, the top-k error ratio is one of the primary methods to measure the accuracy of multi-category classification. Top-k multiclass SVM was designed to minimize the empirical risk based on the top-k error ratio. Two SDCA-based algorithms exist for learning the top-k SVM, both of which have several desirable properties for achieving optimization. However, both algorithms suffer from a serious disadvantage, that is, they cannot attain the optimal convergence in most cases owing to their theoretical imperfections. As demonstrated through numerical simulations, if the modified SDCA algorithm is employed, optimal convergence is always achieved, in contrast to the failure of the two existing SDCA-based algorithms. Finally, our analytical results are presented to clarify the significance of these existing algorithms.
Yoshihiro HIROHASHI
DENSO CORPORATION
Tsuyoshi KATO
Gunma University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yoshihiro HIROHASHI, Tsuyoshi KATO, "Corrected Stochastic Dual Coordinate Ascent for Top-k SVM" in IEICE TRANSACTIONS on Information,
vol. E103-D, no. 11, pp. 2323-2331, November 2020, doi: 10.1587/transinf.2019EDP7261.
Abstract: Currently, the top-k error ratio is one of the primary methods to measure the accuracy of multi-category classification. Top-k multiclass SVM was designed to minimize the empirical risk based on the top-k error ratio. Two SDCA-based algorithms exist for learning the top-k SVM, both of which have several desirable properties for achieving optimization. However, both algorithms suffer from a serious disadvantage, that is, they cannot attain the optimal convergence in most cases owing to their theoretical imperfections. As demonstrated through numerical simulations, if the modified SDCA algorithm is employed, optimal convergence is always achieved, in contrast to the failure of the two existing SDCA-based algorithms. Finally, our analytical results are presented to clarify the significance of these existing algorithms.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2019EDP7261/_p
Copy
@ARTICLE{e103-d_11_2323,
author={Yoshihiro HIROHASHI, Tsuyoshi KATO, },
journal={IEICE TRANSACTIONS on Information},
title={Corrected Stochastic Dual Coordinate Ascent for Top-k SVM},
year={2020},
volume={E103-D},
number={11},
pages={2323-2331},
abstract={Currently, the top-k error ratio is one of the primary methods to measure the accuracy of multi-category classification. Top-k multiclass SVM was designed to minimize the empirical risk based on the top-k error ratio. Two SDCA-based algorithms exist for learning the top-k SVM, both of which have several desirable properties for achieving optimization. However, both algorithms suffer from a serious disadvantage, that is, they cannot attain the optimal convergence in most cases owing to their theoretical imperfections. As demonstrated through numerical simulations, if the modified SDCA algorithm is employed, optimal convergence is always achieved, in contrast to the failure of the two existing SDCA-based algorithms. Finally, our analytical results are presented to clarify the significance of these existing algorithms.},
keywords={},
doi={10.1587/transinf.2019EDP7261},
ISSN={1745-1361},
month={November},}
Copy
TY - JOUR
TI - Corrected Stochastic Dual Coordinate Ascent for Top-k SVM
T2 - IEICE TRANSACTIONS on Information
SP - 2323
EP - 2331
AU - Yoshihiro HIROHASHI
AU - Tsuyoshi KATO
PY - 2020
DO - 10.1587/transinf.2019EDP7261
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E103-D
IS - 11
JA - IEICE TRANSACTIONS on Information
Y1 - November 2020
AB - Currently, the top-k error ratio is one of the primary methods to measure the accuracy of multi-category classification. Top-k multiclass SVM was designed to minimize the empirical risk based on the top-k error ratio. Two SDCA-based algorithms exist for learning the top-k SVM, both of which have several desirable properties for achieving optimization. However, both algorithms suffer from a serious disadvantage, that is, they cannot attain the optimal convergence in most cases owing to their theoretical imperfections. As demonstrated through numerical simulations, if the modified SDCA algorithm is employed, optimal convergence is always achieved, in contrast to the failure of the two existing SDCA-based algorithms. Finally, our analytical results are presented to clarify the significance of these existing algorithms.
ER -