Guardado en:
Detalles Bibliográficos
Autores principales: Hira, Rupkatha, Kau, Dominik, Sorrell, Jessica
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2412.09686
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911536228859904
author Hira, Rupkatha
Kau, Dominik
Sorrell, Jessica
author_facet Hira, Rupkatha
Kau, Dominik
Sorrell, Jessica
contents Active learning aims to reduce the number of labeled data points required by machine learning algorithms by selectively querying labels from initially unlabeled data. Ensuring replicability, where an algorithm produces consistent outcomes across different runs, is essential for the reliability of machine learning models but often increases sample complexity. This paper investigates the cost of replicability in active learning using two classical disagreement-based methods: the CAL and A^2 algorithms. Leveraging randomized thresholding techniques, we propose two replicable active learning algorithms: one for realizable learning of finite hypothesis classes and another for the agnostic setting. Our theoretical analysis shows that while enforcing replicability increases label complexity, CAL and A^2 still achieve substantial label savings under this constraint. These findings provide insights into balancing efficiency and stability in active learning.
format Preprint
id arxiv_https___arxiv_org_abs_2412_09686
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle The Cost of Replicability in Active Learning
Hira, Rupkatha
Kau, Dominik
Sorrell, Jessica
Machine Learning
Active learning aims to reduce the number of labeled data points required by machine learning algorithms by selectively querying labels from initially unlabeled data. Ensuring replicability, where an algorithm produces consistent outcomes across different runs, is essential for the reliability of machine learning models but often increases sample complexity. This paper investigates the cost of replicability in active learning using two classical disagreement-based methods: the CAL and A^2 algorithms. Leveraging randomized thresholding techniques, we propose two replicable active learning algorithms: one for realizable learning of finite hypothesis classes and another for the agnostic setting. Our theoretical analysis shows that while enforcing replicability increases label complexity, CAL and A^2 still achieve substantial label savings under this constraint. These findings provide insights into balancing efficiency and stability in active learning.
title The Cost of Replicability in Active Learning
topic Machine Learning
url https://arxiv.org/abs/2412.09686