Uncertainty Quantification and Resource-Demanding Computer Vision Applications of Deep Learning

TL;DR

This paper reviews methods for uncertainty quantification in deep neural networks, focusing on resource-intensive training techniques for safety-critical applications like autonomous driving and medical imaging.

Contribution

It introduces developed methods for teaching DNNs to recognize uncertainty and learn from few labels, along with resource-demanding neural architecture search strategies.

Findings

Uncertainty quantification improves safety in critical applications.

Training with uncertainty awareness is significantly more resource-intensive.

Neural architecture search adds substantial computational overhead.

Abstract

Bringing deep neural networks (DNNs) into safety critical applications such as automated driving, medical imaging and finance, requires a thorough treatment of the model's uncertainties. Training deep neural networks is already resource demanding and so is also their uncertainty quantification. In this overview article, we survey methods that we developed to teach DNNs to be uncertain when they encounter new object classes. Additionally, we present training methods to learn from only a few labels with help of uncertainty quantification. Note that this is typically paid with a massive overhead in computation of an order of magnitude and more compared to ordinary network training. Finally, we survey our work on neural architecture search which is also an order of magnitude more resource demanding then ordinary network training.