Neural Gaussian Mirror for Controlled Feature Selection in Neural Networks

Abstract

Deep neural networks (DNNs) have become increasingly popular and achieved outstanding performance in predictive tasks. However, the DNN framework itself cannot inform the user which features are more or less relevant for making the prediction, which limits its applicability in many scientific fields. We introduce neural Gaussian mirrors (NGMs), in which mirrored features are created, via a structured perturbation based on a kernel-based conditional dependence measure, to help evaluate feature importance. We design two modifications of the DNN architecture for incorporating mirrored features and providing mirror statistics to measure feature importance. As shown in simulated and real data examples, the proposed method controls the feature selection error rate at a predefined level and maintains a high selection power even with the presence of highly correlated features.

References in corpus (6)

• Dropout Training as Adaptive Regularization
• Veridical Data Science
• DeepPINK: reproducible feature selection in deep neural networks
• Interpretation of Prediction Models Using the Input Gradient
• Interpreting Convolutional Neural Networks Through Compression
• Controlling False Discovery Rate Using Gaussian Mirrors

Cited by in corpus (1)

• Controlling False Discovery Rate Using Gaussian Mirrors