Generalized Uncertainty of Deep Neural Networks: Taxonomy and Applications

TL;DR

This paper reviews and categorizes methods for quantifying and utilizing the uncertainty in deep neural networks, highlighting their importance for interpretability, robustness, and efficiency across various applications.

Contribution

It introduces a generalized definition of uncertainty in deep neural networks and catalogs existing methods and applications, bridging classic uncertainty quantification with deep learning.

Findings

Uncertainty quantification enhances robustness and interpretability.

Generalized uncertainty aids in semi-supervised and weakly-supervised learning.

Methods improve model robustness against noise and adversarial attacks.

Abstract

Deep neural networks have seen enormous success in various real-world applications. Beyond their predictions as point estimates, increasing attention has been focused on quantifying the uncertainty of their predictions. In this review, we show that the uncertainty of deep neural networks is not only important in a sense of interpretability and transparency, but also crucial in further advancing their performance, particularly in learning systems seeking robustness and efficiency. We will generalize the definition of the uncertainty of deep neural networks to any number or vector that is associated with an input or an input-label pair, and catalog existing methods on ``mining'' such uncertainty from a deep model. We will include those methods from the classic field of uncertainty quantification as well as those methods that are specific to deep neural networks. We then show a wide spectrum of applications of such generalized uncertainty in realistic learning tasks including robust learning such as noisy learning, adversarially robust learning; data-efficient learning such as semi-supervised and weakly-supervised learning; and model-efficient learning such as model compression and knowledge distillation.