arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.05336cs.LGstat.ML
收藏

了解实例级标签噪波:不同的影响和处理方法

Understanding Instance-Level Label Noise: Disparate Impacts and Treatments

University of California, Santa Cruz(加州大学圣克鲁兹分校)

Yang Liu

摘要:

本文旨在提供对过参数化模型(例如,深度神经网络)记忆实例相关噪声标签的效果的理解。我们首先量化了记忆噪声实例所造成的危害,并显示了不同的代表频率的样本实例的噪声标签的不同影响。然后,我们分析了几种流行的带噪声标签的学习解决方案如何在实例级别减轻这种伤害。我们的分析表明,现有的方法导致不同的治疗时,处理嘈杂的情况。虽然通过应用这些解决方案,较高频率的实例通常有很高的改善概率,但较低频率的实例却没有。我们的分析揭示了新的理解,当这些方法的工作,并提供了理论依据,以前报道的经验观察。这一观察结果要求我们重新思考标签噪声在实例中的分布,并呼吁对不同制度中的实例进行不同的处理。

英文摘要:

This paper aims to provide understandings for the effect of an over-parameterized model, e.g. a deep neural network, memorizing instance-dependent noisy labels. We first quantify the harms caused by memorizing noisy instances, and show the disparate impacts of noisy labels for sample instances with different representation frequencies. We then analyze how several popular solutions for learning with noisy labels mitigate this harm at the instance level. Our analysis reveals that existing approaches lead to disparate treatments when handling noisy instances. While higher-frequency instances often enjoy a high probability of an improvement by applying these solutions, lower-frequency instances do not. Our analysis reveals new understandings for when these approaches work, and provides theoretical justifications for previously reported empirical observations. This observation requires us to rethink the distribution of label noise across instances and calls for different treatments for instances in different regimes.

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。

↑