Going Beyond One-Hot Encoding in Classification: Can Human Uncertainty Improve Model Performance?

TL;DR

This paper explores integrating human label uncertainty into deep learning models using distributional labels, demonstrating improved generalization, calibration, and performance in remote sensing image classification.

Contribution

It introduces a novel approach to incorporate human uncertainty via distributional labels, enhancing model accuracy and calibration in classification tasks.

Findings

Improved model generalization to unseen data

Better-calibrated probability estimates

Enhanced trustworthiness of predictions

Abstract

Technological and computational advances continuously drive forward the broad field of deep learning. In recent years, the derivation of quantities describing theuncertainty in the prediction - which naturally accompanies the modeling process - has sparked general interest in the deep learning community. Often neglected in the machine learning setting is the human uncertainty that influences numerous labeling processes. As the core of this work, label uncertainty is explicitly embedded into the training process via distributional labels. We demonstrate the effectiveness of our approach on image classification with a remote sensing data set that contains multiple label votes by domain experts for each image: The incorporation of label uncertainty helps the model to generalize better to unseen data and increases model performance. Similar to existing calibration methods, the distributional labels lead to better-calibrated probabilities, which in turn yield more certain and trustworthy predictions.