Deep Neural Networks Under Stress

TL;DR

This paper investigates the robustness of deep neural network features under perturbations, demonstrating their superior resilience and high compression potential with minimal performance loss.

Contribution

It provides an extensive analysis of deep feature resiliency, focusing on the trade-off between compression and accuracy, which was previously underexplored.

Findings

Deep features are more robust to perturbations than classical features.

Achieved 98.4% compression with only 0.88% score loss on Pascal VOC 2007.

Perturbations in feature precision and dimensions impact performance minimally.

Abstract

In recent years, deep architectures have been used for transfer learning with state-of-the-art performance in many datasets. The properties of their features remain, however, largely unstudied under the transfer perspective. In this work, we present an extensive analysis of the resiliency of feature vectors extracted from deep models, with special focus on the trade-off between performance and compression rate. By introducing perturbations to image descriptions extracted from a deep convolutional neural network, we change their precision and number of dimensions, measuring how it affects the final score. We show that deep features are more robust to these disturbances when compared to classical approaches, achieving a compression rate of 98.4%, while losing only 0.88% of their original score for Pascal VOC 2007.