On Uncertainty Calibration for Equivariant Functions

2025-10-24

AI Summary

Systematic miscalibration—evidenced by elevated Expected Calibration Error (ECE) in classification or Expected Normalized Calibration Error (ENCE) in regression—is shown to stem fundamentally from symmetry mismatch between model architecture and data distribution. Method: We establish the first theoretical link between equivariance and uncertainty calibration, deriving tight upper and lower bounds on calibration error under varying group-equivariant structures. Building on this, we propose a calibration framework integrating equivariant neural networks with principled uncertainty estimation. Results: Empirical validation across diverse real-world and synthetic sparse datasets demonstrates that aligning the model’s equivariance group with the inherent symmetries of the data significantly reduces calibration error and enhances reliability—particularly in low-data regimes. Our work provides the first formal bridge between equivariance theory and calibration, revealing symmetry as a dual regulator of both generalization performance and predictive trustworthiness.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Data-sparse settings such as robotic manipulation, molecular physics, and galaxy morphology classification are some of the hardest domains for deep learning. For these problems, equivariant networks can help improve modeling across undersampled parts of the input space, and uncertainty estimation can guard against overconfidence. However, until now, the relationships between equivariance and model confidence, and more generally equivariance and model calibration, has yet to be studied. Since traditional classification and regression error terms show up in the definitions of calibration error, it is natural to suspect that previous work can be used to help understand the relationship between equivariance and calibration error. In this work, we present a theory relating equivariance to uncertainty estimation. By proving lower and upper bounds on uncertainty calibration errors (ECE and ENCE) under various equivariance conditions, we elucidate the generalization limits of equivariant models and illustrate how symmetry mismatch can result in miscalibration in both classification and regression. We complement our theoretical framework with numerical experiments that clarify the relationship between equivariance and uncertainty using a variety of real and simulated datasets, and we comment on trends with symmetry mismatch, group size, and aleatoric and epistemic uncertainties.

Problem

• Studying equivariance's impact on model calibration and uncertainty estimation
• Analyzing how symmetry mismatch causes miscalibration in equivariant models
• Establishing theoretical bounds for uncertainty calibration errors under equivariance

Innovation

• Theoretical bounds on equivariant model calibration errors
• Analyzing symmetry mismatch impact on uncertainty estimation
• Numerical experiments with equivariant networks and uncertainties