Learning with Symmetric Label Noise: The Importance of Being Unhinged
Learning with Symmetric Label Noise: The Importance of Being Unhinged
复制标题
DOI：
--
复制
发表时间：
2015-05
影响因子：
6.6
通讯作者：
Brendan van Rooyen;A. Menon;R. C. Williamson
中科院分区：
文献类型：
--
作者：
Brendan van Rooyen;A. Menon;R. C. Williamson
凸势最小化实际上是二值分类的方法。然而，Long和Servedio[2010]证明，在对称标签噪声（SLN）下，线性函数类上任何凸势的最小化可以导致等同于随机猜测的分类性能。这表面上表明凸损失不是sln鲁棒的。在本文中，我们提出了一个凸的，分类校准的损失，并证明了它是sln鲁棒的。由于损失是负无界的，因此避免了Long and Servedio[2010]的结果。损失是铰链损失的修正，其中不夹紧在零；因此，我们称之为精神错乱的损失。我们证明了最优解等价于强正则化支持向量机的解，并且是任何凸势的极限解；这意味着强l2正则化使得大多数标准学习器具有sln鲁棒性。实验验证了该方法的sln鲁棒性。所以，向王尔德道歉[1895]，虽然真相很少是纯粹的，但它可以是简单的。
Convex potential minimisation is the de facto approach to binary classification. However, Long and Servedio [2010] proved that under symmetric label noise (SLN), minimisation of any convex potential over a linear function class can result in classification performance equivalent to random guessing. This ostensibly shows that convex losses are not SLN-robust. In this paper, we propose a convex, classification-calibrated loss and prove that it is SLN-robust. The loss avoids the Long and Servedio [2010] result by virtue of being negatively unbounded. The loss is a modification of the hinge loss, where one does not clamp at zero; hence, we call it the unhinged loss. We show that the optimal unhinged solution is equivalent to that of a strongly regularised SVM, and is the limiting solution for any convex potential; this implies that strong l2 regularisation makes most standard learners SLN-robust. Experiments confirm the unhinged loss' SLN-robustness is borne out in practice. So, with apologies to Wilde [1895], while the truth is rarely pure, it can be simple.