We describe a simple and effective method (Spectral Attribute removaL; SAL) to remove private or guarded information from neural representations. Our method uses matrix decomposition to project the input representations into directions with reduced covariance with the guarded information rather than maximal covariance as factorization methods normally use. We begin with linear information removal and proceed to generalize our algorithm to the case of nonlinear information removal using kernels. Our experiments demonstrate that our algorithm retains better main task performance after removing the guarded information compared to previous work. In addition, our experiments demonstrate that we need a relatively small amount of guarded attribute data to remove information about these attributes, which lowers the exposure to sensitive data and is more suitable for low-resource scenarios. Code is available at https://github.com/jasonshaoshun/SAL.
翻译:我们描述一种简单而有效的方法(Spetratracal removaL; SAL),将私人或保密信息从神经显示器中去除。我们的方法使用矩阵分解法,将输入表达器投射到与保密信息相对较少的共差,而不是最大共差,作为通常使用的系数化方法。我们首先从线性信息删除开始,然后将我们的算法概括到使用内核去除非线性信息的情况。我们的实验表明,我们的算法在去除保密信息后保持了比以往工作更好的主要任务性能。此外,我们的实验还表明,我们需要相对较少的保密属性数据来删除关于这些属性的信息,从而降低敏感数据的暴露度,更适合低资源情景。代码可在https://github.com/sonshaoshun/SAL查阅。