深度学习中的可能/可能类别 [英] Possible/maybe category in deep learning
问题描述
我有兴趣利用我在深度学习任务中拥有的一些部分标记的数据.我使用的是完全卷积的方法,而不是从标记区域中采样补丁.
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.
我有一些掩码可以勾勒出图像中明确正例的区域,但图像中未掩蔽的区域不一定是负的——它们可能是正的.有谁知道将此类课程纳入深度学习环境的方法?
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?
Triplet/contrastive loss 似乎是要走的路,但我不确定如何适应模糊"或模糊的负/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,如深度学习书籍:
我们可以假设对于一些小的常数 eps
,训练集标签 y 是正确的,概率为 1 - eps
,否则任何其他可能的标签都可能是正确.
We can assume that for some small constant
eps
, the training set label y is correct with probability1 - eps
, and otherwise any of the other possible labels might be correct.
标签平滑基于带有 k
个输出值的 softmax 正则化模型,方法是用 eps/k 和 1 - (k - 1)/的目标替换硬 0 和 1 分类目标k * eps
,分别.
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.
查看我关于在 Pandas 中实现标签平滑的问题.
否则,如果您确定某些领域是负面的,有些领域是正面的,而有些领域是不确定的,那么您可以引入第三个不确定类.我处理过包含不确定类的数据集,这些类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/可能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!