GIT: Detecting Uncertainty, Out-Of-Distribution and Adversarial Samples using Gradients and Invariance Transformations

TL;DR

GIT is a comprehensive method that detects neural network errors by combining gradient analysis and invariance transformations to identify out-of-distribution, adversarial, and misclassified samples.

Contribution

It introduces a novel holistic approach that integrates gradient information with invariance transformations for improved error detection in neural networks.

Findings

GIT outperforms state-of-the-art methods across various architectures.

It effectively detects out-of-distribution and adversarial samples.

The approach generalizes well to different problem setups.

Abstract

Deep neural networks tend to make overconfident predictions and often require additional detectors for misclassifications, particularly for safety-critical applications. Existing detection methods usually only focus on adversarial attacks or out-of-distribution samples as reasons for false predictions. However, generalization errors occur due to diverse reasons often related to poorly learning relevant invariances. We therefore propose GIT, a holistic approach for the detection of generalization errors that combines the usage of gradient information and invariance transformations. The invariance transformations are designed to shift misclassified samples back into the generalization area of the neural network, while the gradient information measures the contradiction between the initial prediction and the corresponding inherent computations of the neural network using the transformed sample. Our experiments demonstrate the superior performance of GIT compared to the state-of-the-art on a variety of network architectures, problem setups and perturbation types.