Calibration in Machine Learning Uncertainty Quantification: beyond consistency to target adaptivity
This work addresses the need for more robust uncertainty quantification in ML for materials and chemical science, though it is incremental by building on existing calibration concepts.
The paper tackles the problem that average calibration and consistency are insufficient for reliable uncertainty quantification in machine learning regression, showing that consistency does not guarantee adaptivity (calibration with respect to input features), and proposes validation methods to address this gap.
Reliable uncertainty quantification (UQ) in machine learning (ML) regression tasks is becoming the focus of many studies in materials and chemical science. It is now well understood that average calibration is insufficient, and most studies implement additional methods testing the conditional calibration with respect to uncertainty, i.e. consistency. Consistency is assessed mostly by so-called reliability diagrams. There exists however another way beyond average calibration, which is conditional calibration with respect to input features, i.e. adaptivity. In practice, adaptivity is the main concern of the final users of a ML-UQ method, seeking for the reliability of predictions and uncertainties for any point in features space. This article aims to show that consistency and adaptivity are complementary validation targets, and that a good consistency does not imply a good adaptivity. Adapted validation methods are proposed and illustrated on a representative example.