Agreeing to disagree: active learning with noisy labels without crowdsourcing

We propose a new active learning method for classification, which handles label noise without relying on multiple oracles (i.e., crowdsourcing). We propose a strategy that selects (for labeling) instances with a high influence on the learned model. An instance x is said to have a high influence on t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:International journal of machine learning and cybernetics Jg. 9; H. 8; S. 1307 - 1319
Hauptverfasser: Bouguelia, Mohamed-Rafik, Nowaczyk, Slawomir, Santosh, K. C., Verikas, Antanas
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Berlin/Heidelberg Springer Berlin Heidelberg 01.08.2018
Springer Nature B.V
Schlagworte:
Active learning
Algorithms
Artificial Intelligence
Classification
Complex Systems
Computational Intelligence
Control
Crowdsourcing
Data mining
Datasets
Engineering
Interactive learning
Label noise
Labeling
Labels
Learning
Machine learning
Mechatronics
Mislabeling
Original Article
Pattern Recognition
Robotics
Systems Biology
Teaching methods
Training
Mislabeling
Active learning
Label noise
Classification
ISSN:1868-8071, 1868-808X, 1868-808X
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We propose a new active learning method for classification, which handles label noise without relying on multiple oracles (i.e., crowdsourcing). We propose a strategy that selects (for labeling) instances with a high influence on the learned model. An instance x is said to have a high influence on the model h , if training h on x (with label y = h ( x ) ) would result in a model that greatly disagrees with h on labeling other instances. Then, we propose another strategy that selects (for labeling) instances that are highly influenced by changes in the learned model. An instance x is said to be highly influenced, if training h with a set of instances would result in a committee of models that agree on a common label for x but disagree with h ( x ). We compare the two strategies and we show, on different publicly available datasets, that selecting instances according to the first strategy while eliminating noisy labels according to the second strategy, greatly improves the accuracy compared to several benchmarking methods, even when a significant amount of instances are mislabeled.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1868-8071
1868-808X
1868-808X
DOI:10.1007/s13042-017-0645-0