arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2106.09643cs.AI
收藏

MetaBalance:用于类不平衡数据的高性能神经网络

MetaBalance: High-Performance Neural Networks for Class-Imbalanced Data

University of Maryland, College Park(马里兰大学帕克分校);Capital One(第一资本金融公司);Center for Machine Learning(机器学习中心)

Arpit Bansal, Micah Goldblum, Valeriia Cherepanova, Avi Schwarzschild, C. Bayan Bruss, Tom Goldstein

摘要:

类不平衡数据,其中一些类包含的样本比其他的多得多,在现实世界的应用中是普遍存在的。用于处理类不平衡的标准技术通常通过在重新加权的损失或重新平衡的数据上进行训练来工作。不幸的是,在这样的目标上训练过度参数化的神经网络会导致少数类数据的快速记忆。为了避免这个陷阱,我们利用元学习,它使用“外环”和“内环”损失,每一个都可以使用不同的策略来平衡。我们评估我们的方法,MetaBalance,图像分类,信用卡欺诈检测,贷款违约预测和面部识别任务与严重不平衡的数据,我们发现MetaBalance优于广泛的流行的重采样策略。

英文摘要:

Class-imbalanced data, in which some classes contain far more samples than others, is ubiquitous in real-world applications. Standard techniques for handling class-imbalance usually work by training on a re-weighted loss or on re-balanced data. Unfortunately, training overparameterized neural networks on such objectives causes rapid memorization of minority class data. To avoid this trap, we harness meta-learning, which uses both an ''outer-loop'' and an ''inner-loop'' loss, each of which may be balanced using different strategies. We evaluate our method, MetaBalance, on image classification, credit-card fraud detection, loan default prediction, and facial recognition tasks with severely imbalanced data, and we find that MetaBalance outperforms a wide array of popular re-sampling strategies.

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。

↑