Partial Label Metric Learning Based on Statistical Inference
Summary :
In partial label data, the ground-truth label of a training example is concealed in a set of candidate labels associated with the instance. As the ground-truth label is inaccessible, it is difficult to train the classifier via the label information. Consequently, manifold structure information is adopted, which is under the assumption that neighbor/similar instances in the feature space have similar labels in the label space. However, the real-world data may not fully satisfy this assumption. In this paper, a partial label metric learning method based on likelihood-ratio test is proposed to make partial label data satisfy the manifold assumption. Moreover, the proposed method needs no objective function and treats the data pairs asymmetrically. The experimental results on several real-world PLL datasets indicate that the proposed method outperforms the existing partial label metric learning methods in terms of classification accuracy and disambiguation accuracy while costs less time.
- Publication
- IEICE TRANSACTIONS on Information Vol.E103-D No.6 pp.1355-1361
- Publication Date
- 2020/06/01
- Publicized
- 2020/03/05
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.2019EDP7182
- Type of Manuscript
- PAPER
- Category
- Artificial Intelligence, Data Mining
Authors
Tian XIE
Information Engineering University
Hongchang CHEN
China National Digital Switching System Engineering & Technological R&D Center
Tuosiyu MING
China National Digital Switching System Engineering & Technological R&D Center
Jianpeng ZHANG
China National Digital Switching System Engineering & Technological R&D Center
Chao GAO
China National Digital Switching System Engineering & Technological R&D Center
Shaomei LI
China National Digital Switching System Engineering & Technological R&D Center
Yuehang DING
China National Digital Switching System Engineering & Technological R&D Center
Keyword
Latest Issue
Copyrights notice of machine-translated contents
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.
Cite this
Copy
Tian XIE, Hongchang CHEN, Tuosiyu MING, Jianpeng ZHANG, Chao GAO, Shaomei LI, Yuehang DING, "Partial Label Metric Learning Based on Statistical Inference" in IEICE TRANSACTIONS on Information,
vol. E103-D, no. 6, pp. 1355-1361, June 2020, doi: 10.1587/transinf.2019EDP7182.
Abstract: In partial label data, the ground-truth label of a training example is concealed in a set of candidate labels associated with the instance. As the ground-truth label is inaccessible, it is difficult to train the classifier via the label information. Consequently, manifold structure information is adopted, which is under the assumption that neighbor/similar instances in the feature space have similar labels in the label space. However, the real-world data may not fully satisfy this assumption. In this paper, a partial label metric learning method based on likelihood-ratio test is proposed to make partial label data satisfy the manifold assumption. Moreover, the proposed method needs no objective function and treats the data pairs asymmetrically. The experimental results on several real-world PLL datasets indicate that the proposed method outperforms the existing partial label metric learning methods in terms of classification accuracy and disambiguation accuracy while costs less time.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2019EDP7182/_p
Copy
@ARTICLE{e103-d_6_1355,
author={Tian XIE, Hongchang CHEN, Tuosiyu MING, Jianpeng ZHANG, Chao GAO, Shaomei LI, Yuehang DING, },
journal={IEICE TRANSACTIONS on Information},
title={Partial Label Metric Learning Based on Statistical Inference},
year={2020},
volume={E103-D},
number={6},
pages={1355-1361},
abstract={In partial label data, the ground-truth label of a training example is concealed in a set of candidate labels associated with the instance. As the ground-truth label is inaccessible, it is difficult to train the classifier via the label information. Consequently, manifold structure information is adopted, which is under the assumption that neighbor/similar instances in the feature space have similar labels in the label space. However, the real-world data may not fully satisfy this assumption. In this paper, a partial label metric learning method based on likelihood-ratio test is proposed to make partial label data satisfy the manifold assumption. Moreover, the proposed method needs no objective function and treats the data pairs asymmetrically. The experimental results on several real-world PLL datasets indicate that the proposed method outperforms the existing partial label metric learning methods in terms of classification accuracy and disambiguation accuracy while costs less time.},
keywords={},
doi={10.1587/transinf.2019EDP7182},
ISSN={1745-1361},
month={June},}
Copy
TY - JOUR
TI - Partial Label Metric Learning Based on Statistical Inference
T2 - IEICE TRANSACTIONS on Information
SP - 1355
EP - 1361
AU - Tian XIE
AU - Hongchang CHEN
AU - Tuosiyu MING
AU - Jianpeng ZHANG
AU - Chao GAO
AU - Shaomei LI
AU - Yuehang DING
PY - 2020
DO - 10.1587/transinf.2019EDP7182
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E103-D
IS - 6
JA - IEICE TRANSACTIONS on Information
Y1 - June 2020
AB - In partial label data, the ground-truth label of a training example is concealed in a set of candidate labels associated with the instance. As the ground-truth label is inaccessible, it is difficult to train the classifier via the label information. Consequently, manifold structure information is adopted, which is under the assumption that neighbor/similar instances in the feature space have similar labels in the label space. However, the real-world data may not fully satisfy this assumption. In this paper, a partial label metric learning method based on likelihood-ratio test is proposed to make partial label data satisfy the manifold assumption. Moreover, the proposed method needs no objective function and treats the data pairs asymmetrically. The experimental results on several real-world PLL datasets indicate that the proposed method outperforms the existing partial label metric learning methods in terms of classification accuracy and disambiguation accuracy while costs less time.
ER -
English
