When Optimizing f-divergence is Robust with Label Noise
When Optimizing f-divergence is Robust with Label Noise
复制标题
DOI：
--
复制
发表时间：
2020-11
期刊：
影响因子：
--
通讯作者：
Jiaheng Wei;Yang Liu
中科院分区：
文献类型：
--
作者：
Jiaheng Wei;Yang Liu
我们展示了当最大化一个适当定义的关于分类器预测的$f$散度度量时，监督标签具有标签噪声的鲁棒性。利用其变分形式，当标签噪声存在时，我们为这个特定的$f$-散度导出了一个很好的解耦性质，其中散度显示为在干净分布上定义的变分差和由于噪声引入的偏置项的线性组合。上述推导有助于我们分析该度量对于不同的f -散度函数的鲁棒性。有了已建立的鲁棒性，这组$f$散度函数可以作为有噪声标签学习问题的有用度量，不需要指定标签的噪声率。当它们可能不健壮时，我们会提出修复方案。除了分析结果外，我们还提出了全面的实验研究。
We show when maximizing a properly defined $f$-divergence measure with respect to a classifier's predictions and the supervised labels is robust with label noise. Leveraging its variational form, we derive a nice decoupling property for this particular $f$-divergence when label noise presents, where the divergence is shown to be a linear combination of the variational difference defined on the clean distribution and a bias term introduced due to the noise. The above derivation helps us analyze the robustness of this measure for different $f$-divergence functions. With established robustness, this family of $f$-divergence functions arises as useful metrics for the problem of learning with noisy labels, which do not require the specification of the labels' noise rate. When they are possibly not robust, we propose fixes to make them so. In addition to the analytical results, we present thorough experimental studies.