On the Ramifications of Human Label Uncertainty

TL;DR

This paper investigates how human label disagreement affects uncertainty estimation in machine learning, revealing limitations of current methods and proposing a novel NSS-based label dilution approach to mitigate these effects.

Contribution

It introduces a new NSS-based label dilution training scheme that reduces the impact of human label uncertainty without requiring extensive human labels.

Findings

Existing uncertainty metrics are limited under human label uncertainty.

NSS-based label dilution improves model robustness to label disagreement.

Training with diluted labels alleviates undue effects of human label uncertainty.

Abstract

Humans exhibit disagreement during data labeling. We term this disagreement as human label uncertainty. In this work, we study the ramifications of human label uncertainty (HLU). Our evaluation of existing uncertainty estimation algorithms, with the presence of HLU, indicates the limitations of existing uncertainty metrics and algorithms themselves in response to HLU. Meanwhile, we observe undue effects in predictive uncertainty and generalizability. To mitigate the undue effects, we introduce a novel natural scene statistics (NSS) based label dilution training scheme without requiring massive human labels. Specifically, we first select a subset of samples with low perceptual quality ranked by statistical regularities of images. We then assign separate labels to each sample in this subset to obtain a training set with diluted labels. Our experiments and analysis demonstrate that training with NSS-based label dilution alleviates the undue effects caused by HLU.