arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.12844cs.LG
收藏

利用广义对抗距离有效发现分类器错误

Generalized Adversarial Distances to Efficiently Discover Classifier Errors

Air Force Research Lab(空军研究实验室);Miami University(迈阿密大学);General Mills(通用磨坊公司)

Walter Bennette, Sally Dufek, Karsten Maurer, Sean Sisti, Bunyod Tusmatov

摘要:

给定一个黑盒分类模型和一个来自某个应用领域的未标记评估数据集,需要开发有效的策略来评估模型。随机抽样允许用户估计准确度、精确度和召回率等指标,但可能无法提供对高置信度错误的洞察。高置信度错误是模型对其预测非常有信心,但却是错误的罕见事件。这些错误可能代表代价高昂的错误,应该明确搜索。在本文中,我们提出了一种对对抗距离搜索的推广,该搜索利用对抗机器学习的概念来识别分类器可能过于自信的预测。在寻找高置信度错误时,这些预测是有用的样本实例,因为它们的错误率往往高于预期。我们的泛化允许对抗距离应用于任何分类器或数据域。实验结果表明,广义方法发现的错误率大于预期的抽样预测的信心,并优于竞争的方法。

英文摘要:

Given a black-box classification model and an unlabeled evaluation dataset from some application domain, efficient strategies need to be developed to evaluate the model. Random sampling allows a user to estimate metrics like accuracy, precision, and recall, but may not provide insight to high-confidence errors. High-confidence errors are rare events for which the model is highly confident in its prediction, but is wrong. Such errors can represent costly mistakes and should be explicitly searched for. In this paper we propose a generalization to the Adversarial Distance search that leverages concepts from adversarial machine learning to identify predictions for which a classifier may be overly confident. These predictions are useful instances to sample when looking for high-confidence errors because they are prone to a higher rate of error than expected. Our generalization allows Adversarial Distance to be applied to any classifier or data domain. Experimental results show that the generalized method finds errors at rates greater than expected given the confidence of the sampled predictions, and outperforms competing methods.

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。

↑