The Aleatoric Uncertainty Estimation Using a Separate Formulation with Virtual Residuals

TL;DR

This paper introduces a novel optimization framework for aleatoric uncertainty estimation in regression tasks, effectively separating signal and uncertainty estimation to improve accuracy and avoid overfitting.

Contribution

It proposes a new separable formulation with virtual residuals for more accurate aleatoric uncertainty estimation, outperforming existing methods.

Findings

Outperforms state-of-the-art techniques in uncertainty estimation

Effective in regression, age, and depth estimation tasks

Reduces underestimation of predictive uncertainty

Abstract

We propose a new optimization framework for aleatoric uncertainty estimation in regression problems. Existing methods can quantify the error in the target estimation, but they tend to underestimate it. To obtain the predictive uncertainty inherent in an observation, we propose a new separable formulation for the estimation of a signal and of its uncertainty, avoiding the effect of overfitting. By decoupling target estimation and uncertainty estimation, we also control the balance between signal estimation and uncertainty estimation. We conduct three types of experiments: regression with simulation data, age estimation, and depth estimation. We demonstrate that the proposed method outperforms a state-of-the-art technique for signal and uncertainty estimation.