A Forward-Backward Approach for Visualizing Information Flow in Deep Networks

TL;DR

This paper presents a new systematic and efficient framework for visualizing information flow in deep convolutional networks, highlighting the features contributing to specific model explanations.

Contribution

The proposed method offers a robust, high-resolution visualization of features in deep networks, improving over existing visualization techniques.

Findings

Supports the benefits of the framework over existing methods

Produces high-resolution, compact support visualizations

Efficient and numerically robust

Abstract

We introduce a new, systematic framework for visualizing information flow in deep networks. Specifically, given any trained deep convolutional network model and a given test image, our method produces a compact support in the image domain that corresponds to a (high-resolution) feature that contributes to the given explanation. Our method is both computationally efficient as well as numerically robust. We present several preliminary numerical results that support the benefits of our framework over existing methods.