Regression with re-labeling for noisy data

Youngdoo Son, Seokho Kang

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Active learning, which focuses on building an accurate prediction model with a reduced cost by actively querying which instances should be labeled for training, has been successfully employed in several real-world applications involving expensive labeling costs. Although most existing active learning strategies have focused on labeling unlabeled instances, it has been shown that improving the quality of previously annotated labels is also important when the annotator produces noisy labels. In this study, we propose a novel active learning framework for regression, which is effective for the scenarios with noisy annotators, by providing a new sampling strategy named exploration-refinement (ER) sampling. The ER sampling performs two main steps: exploration and refinement. The exploration step involves finding unlabeled instances to be labeled, and the refinement step seeks to improve the accuracy of already labeled instances. The experimental results on several benchmark datasets demonstrate the effectiveness of the ER sampling with statistical significance.

Original languageEnglish
Pages (from-to)578-587
Number of pages10
JournalExpert Systems with Applications
Volume114
DOIs
StatePublished - 30 Dec 2018

Keywords

  • Active learning
  • Exploration-refinement sampling
  • Re-labeling
  • Regression

Fingerprint

Dive into the research topics of 'Regression with re-labeling for noisy data'. Together they form a unique fingerprint.

Cite this