About this document
DeepROCK: Controlled Interaction Detection by nescaufervente is a document available to read on EtoBox.
The paper introduces DeepROCK, a method for error-controlled interaction detection in deep neural networks (DNNs) that enhances interpretability by systematically prioritizing feature interactions while controlling the false discovery rate (FDR). It utilizes a novel DNN architecture with a pairwise-coupling layer and knockoff features to improve statistical power and robustness against noise. Extensive experiments validate DeepROCK
- Author
- nescaufervente
- Language
- EN