深度学习中的可能/也许类别 [英] Possible/maybe category in deep learning

问题描述

我对利用我在深度学习任务中拥有的一些部分标记的数据感兴趣.我使用的是完全卷积方法，而不是从标记区域采样补丁.

I'm interested in taking advantage of some partially labeled data that I have in a deep learning task. I'm using a fully convolutional approach, not sampling patches from the labeled regions.

我有一些遮罩，这些遮罩概述了图像中肯定的正例的区域，但是图像中未​​遮罩的区域不一定是负的-它们可能是正的.有谁知道将这种类型的课程纳入深度学习环境的方法吗?

I have masks that outline regions of definite positive examples in an image, but the unmasked regions in the images are not necessarily negative - they may be positive. Does anyone know of a way to incorporate this type of class in a deep learning setting?

三重/相反的损失似乎是可行的方法，但是我不确定如何适应模糊"或模棱两可的负/正空间.

Triplet/contrastive loss seems like it may be the way to go, but I'm not sure how to accommodate the "fuzzy" or ambiguous negative/positive space.

推荐答案

尝试标签平滑，如

我们可以假设对于某个小的常数eps，训练集标签y以概率1 - eps是正确的，否则，其他任何可能的标签都可能是正确的.

We can assume that for some small constant eps, the training set label y is correct with probability 1 - eps, and otherwise any of the other possible labels might be correct.

标签平滑通过将硬0和1分类目标分别替换为目标eps / k and 1 - (k - 1) / k * eps，从而基于带有k输出值的softmax对模型进行正则化.

Label smoothing regularizes a model based on a softmax with k output values by replacing the hard 0 and 1 classification targets with targets of eps / k and 1 - (k - 1) / k * eps, respectively.

请参阅有关在熊猫中实现标签平滑的问题.

否则，如果您确定知道某些区域是负面的，其他区域是正面的而某些区域是不确定的，则可以引入第三个不确定类.我已经处理了包含 unertain 类的数据集，这些数据集对应于可能属于任何可用类的样本.

Otherwise if you know for sure, that some areas are negative, other are positive while some are uncertain, then you can introduce a third uncertain class. I have worked with data sets that contained uncertain class, which corresponded to samples that could belong to any of the available classes.

这篇关于深度学习中的可能/也许类别的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！