Success of Uncertainty-Aware Deep Models Depends on Data Manifold Geometry

TL;DR

This paper evaluates six uncertainty-aware deep learning models on edge-case tasks, revealing that the data manifold's geometry significantly influences model success, guiding future model selection and development.

Contribution

It demonstrates the importance of data manifold geometry in the effectiveness of uncertainty-aware models for edge-case detection.

Findings

Model performance varies with data manifold geometry

Geometry influences robustness to adversarial attacks

Guides future model selection based on data structure

Abstract

For responsible decision making in safety-critical settings, machine learning models must effectively detect and process edge-case data. Although existing works show that predictive uncertainty is useful for these tasks, it is not evident from literature which uncertainty-aware models are best suited for a given dataset. Thus, we compare six uncertainty-aware deep learning models on a set of edge-case tasks: robustness to adversarial attacks as well as out-of-distribution and adversarial detection. We find that the geometry of the data sub-manifold is an important factor in determining the success of various models. Our finding suggests an interesting direction in the study of uncertainty-aware deep learning models.