Controllable Invariance through Adversarial Feature Learning 1705.11122.pdf

Given paired observations <x,y>, we are interested in the task of predicting the target y based on thvalue of x using a discriminative approach, i.e. directly modeling the conditional distribution p(y|x).As the input x can have highly complicated structure, we employ a dedicated model or algorithm toextract an expressive representation h from x. In addition, we have access to some intrinsic attribute sof x as well as a prior belief that the prediction result should be invariant to s. Thus, when we extractthe representation h from x, we want the representation h to preserve variations that are necessary topredict y while eliminating information of s.
给定成对观测值<x,y>，我们感兴趣的是使用判别方法基于x的值来预测目标y的任务，即直接建模条件分布p(y|x)。由于输入x可以具有高度复杂的结构 我们采用一种专用的模型或算法从x中提取一个表达式h。 此外,我们可以访问一些x的内在属性以及先验的观点，即预测结果应该是不变的。 因此，当我们从x中提取表示h时，我们希望表示h在消除s的信息时保留必要的预测y的变化。

To achieve the aforementioned goal, we employ a deterministic encoder E to obtain the representation by encoding x and s into h, namely, h = E(x, s). It should be noted here that we are using s as an additional input. Intuitively, this can inform and guide the encoder to remove information about undesired variations within the representation. For example, if we want to learn a representation of image x that is invariant to the lighting condition s, the model can learn to “brighten” the representation if it knows the original picture is dark, and vice versa.
为了实现上述目标，我们采用确定性编码器E来通过将x和s编码为h来获得表示，即h = E（x，s）。 这里应该注意，我们使用s作为附加输入。 直观地，这可以通知和指导编码器去除表示内的不期望的变化的信息。 例如，如果我们要学习对照明条件s不变的图像x的表示，则如果知道原始图像较暗，模型可以学习“增亮”表示，反之亦然。