arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2412.09686cs.LG
收藏

主动学习中可重复性的成本

The Cost of Replicability in Active Learning

Johns Hopkins University(约翰霍普金斯大学);University of Pennsylvania(宾夕法尼亚大学);Penn Institute for Computational Science(宾夕法尼亚大学计算科学研究所)

Rupkatha Hira, Dominik Kau, Jessica Sorrell

AI总结:

本文研究了主动学习中可重复性带来的样本复杂性增加,提出两种可重复的主动学习算法,在保证一致性的同时实现标签节省。

摘要:

主动学习旨在通过选择性地查询初始未标记数据的标签来减少机器学习算法所需的标记数据点数量。确保可重复性,即算法在不同运行中产生一致的结果,对于机器学习模型的可靠性至关重要,但通常会增加样本复杂性。本文利用随机阈值技术,提出两种可重复的主动学习算法:一种用于有限假设类的可实现学习,另一种用于非可实现设置。我们的理论分析表明,尽管强制可重复性会增加标签复杂性,CAL和A²算法在该约束下仍能实现显著的标签节省。这些发现为平衡主动学习中的效率和稳定性提供了见解。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑