论文标题
深神经网络的一般分位数损失
Generalized Quantile Loss for Deep Neural Networks
论文作者
论文摘要
本说明提供了一种在回归神经网中添加计数（或分位数）约束的简单方法，因此在培训集中给定$ n $样本的情况下，它保证了$ m <n $样本的预测将大于实际值（标签）。与标准的分位数回归网络不同，提出的方法可以应用于任何损耗函数，而不一定应用于标准的分位数回归损耗，这可以最大程度地减少平均绝对差异。由于该计数约束几乎到处都有零梯度，因此无法使用标准梯度下降方法对其进行优化。为了克服这个问题，通过一些理论分析介绍了基于标准神经网络优化程序的交替方案。
This note presents a simple way to add a count (or quantile) constraint to a regression neural net, such that given $n$ samples in the training set it guarantees that the prediction of $m<n$ samples will be larger than the actual value (the label). Unlike standard quantile regression networks, the presented method can be applied to any loss function and not necessarily to the standard quantile regression loss, which minimizes the mean absolute differences. Since this count constraint has zero gradients almost everywhere, it cannot be optimized using standard gradient descent methods. To overcome this problem, an alternation scheme, which is based on standard neural network optimization procedures, is presented with some theoretical analysis.
