A Data-Driven Measure of Relative Uncertainty for Misclassification Detection

Eduardo Dadalto Câmara Gomes, Marco Romanelli, Georg Pichler, Pablo Piantanida

Keywords: Mathematical methods for Misclassification detection, Uncertainty estimation, Trustworthy AI, Safety

TL;DR: We introduce a data-driven measure of relative uncertainty as a new method for detecting samples misclassified by machine learning classification models.

Abstract: Misclassification detection is an important problem in machine learning, as it allows for the identification of instances where the model's predictions are unreliable. However, conventional uncertainty measures such as Shannon entropy do not provide an effective way to infer the real uncertainty associated with the model's predictions. In this paper, we introduce a novel data-driven measure of relative uncertainty to an observer for misclassification detection. By learning patterns in the distribution of soft-predictions, our uncertainty measure can identify misclassified samples based on the predicted class probabilities. Interestingly, according to the proposed measure, soft-predictions that correspond to misclassified instances can carry a large amount of uncertainty, even though they may have low Shannon entropy. We demonstrate empirical improvements over multiple image classification tasks, outperforming state-of-the-art misclassification detection methods.

Submission Number: 9