Is Uncertainty Quantification a Viable Alternative to Learned Deferral?

TL;DR

This study compares uncertainty quantification methods and learned deferral models in AI for ophthalmology, finding that uncertainty methods may offer more robustness to data shifts and out-of-distribution cases.

Contribution

The paper provides an extensive evaluation of uncertainty quantification versus learned deferral models in clinical AI, highlighting the potential robustness of uncertainty methods in out-of-distribution scenarios.

Findings

Uncertainty quantification methods show promising results for AI deferral.

Models based on uncertainty are more robust to out-of-distribution data.

Evaluation conducted on a large ophthalmology dataset for glaucoma detection.

Abstract

Artificial Intelligence (AI) holds the potential to dramatically improve patient care. However, it is not infallible, necessitating human-AI-collaboration to ensure safe implementation. One aspect of AI safety is the models' ability to defer decisions to a human expert when they are likely to misclassify autonomously. Recent research has focused on methods that learn to defer by optimising a surrogate loss function that finds the optimal trade-off between predicting a class label or deferring. However, during clinical translation, models often face challenges such as data shift. Uncertainty quantification methods aim to estimate a model's confidence in its predictions. However, they may also be used as a deferral strategy which does not rely on learning from specific training distribution. We hypothesise that models developed to quantify uncertainty are more robust to out-of-distribution (OOD) input than learned deferral models that have been trained in a supervised fashion. To investigate this hypothesis, we constructed an extensive evaluation study on a large ophthalmology dataset, examining both learned deferral models and established uncertainty quantification methods, assessing their performance in- and out-of-distribution. Specifically, we evaluate their ability to accurately classify glaucoma from fundus images while deferring cases with a high likelihood of error. We find that uncertainty quantification methods may be a promising choice for AI deferral.