Overcoming Labeling Ability for Latent Positives: Automatic Label Correction along Data Series

Azusa Sawada, Takashi Shibata

2019

Abstract

Although recent progress in machine learning has substantially improved the accuracy of pattern recognition and classification task, the performances of these learned models depend on the annotation quality. Therefore, in the real world, the accuracy of these models is limited by the labelling skills of the annotators. To tackle this problem, we propose a novel learning framework that can obtain an accurate model by finding latent positive samples that are often overlooked by non-skilled annotators. The key of the proposed method is to focus on the data series that is helpful to find the latent positive labels. The proposed method has two main interacting components: 1) a label correction part to seek positives along data series and 2) a model training part on modified labels. The experimental results on simulated data show that the proposed method can obtain the same performance as supervision by oracle label and outperforms the existing method in terms of area under the curve (AUC).

Paper Citation

in Harvard Style

Sawada A. and Shibata T. (2019). Overcoming Labeling Ability for Latent Positives: Automatic Label Correction along Data Series.In Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-351-3, pages 406-413. DOI: 10.5220/0007341704060413

in Bibtex Style

@conference{icpram19,
author={Azusa Sawada and Takashi Shibata},
title={Overcoming Labeling Ability for Latent Positives: Automatic Label Correction along Data Series},
booktitle={Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2019},
pages={406-413},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007341704060413},
isbn={978-989-758-351-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Overcoming Labeling Ability for Latent Positives: Automatic Label Correction along Data Series
SN - 978-989-758-351-3
AU - Sawada A.
AU - Shibata T.
PY - 2019
SP - 406
EP - 413
DO - 10.5220/0007341704060413