A Symmetric Loss Perspective of Reliable Machine Learning
This work provides an overview of symmetric losses and their applications, which is useful for researchers and practitioners in machine learning seeking robust classification methods, particularly when dealing with noisy or corrupted labels. It is an incremental review.
This paper reviews symmetric losses, which are surrogate losses for binary classification that satisfy a specific symmetric condition. It demonstrates their utility in achieving robust classification from corrupted labels in balanced error rate minimization and area under the receiver operating characteristic curve (AUC) maximization, and shows how robust AUC maximization can be applied to natural language processing tasks involving learning from relevant keywords and unlabeled documents.
When minimizing the empirical risk in binary classification, it is a common practice to replace the zero-one loss with a surrogate loss to make the learning objective feasible to optimize. Examples of well-known surrogate losses for binary classification include the logistic loss, hinge loss, and sigmoid loss. It is known that the choice of a surrogate loss can highly influence the performance of the trained classifier and therefore it should be carefully chosen. Recently, surrogate losses that satisfy a certain symmetric condition (aka., symmetric losses) have demonstrated their usefulness in learning from corrupted labels. In this article, we provide an overview of symmetric losses and their applications. First, we review how a symmetric loss can yield robust classification from corrupted labels in balanced error rate (BER) minimization and area under the receiver operating characteristic curve (AUC) maximization. Then, we demonstrate how the robust AUC maximization method can benefit natural language processing in the problem where we want to learn only from relevant keywords and unlabeled documents. Finally, we conclude this article by discussing future directions, including potential applications of symmetric losses for reliable machine learning and the design of non-symmetric losses that can benefit from the symmetric condition.