Conformalized Exceptional Model Mining: Telling Where Your Model Performs (Not) Well

TL;DR

This paper presents Conformalized Exceptional Model Mining, a framework that identifies data subgroups where models perform exceptionally well or poorly, improving interpretability and uncertainty quantification in high-stakes AI applications.

Contribution

It introduces a novel framework combining Conformal Prediction with Exceptional Model Mining to detect and explain regions of model performance deviation.

Findings

Effectively uncovers interpretable subgroups with distinct performance patterns.

Provides rigorous uncertainty quantification with conformal guarantees.

Demonstrates applicability across diverse datasets and tasks.

Abstract

Understanding the nuanced performance of machine learning models is essential for responsible deployment, especially in high-stakes domains like healthcare and finance. This paper introduces a novel framework, Conformalized Exceptional Model Mining, which combines the rigor of Conformal Prediction with the explanatory power of Exceptional Model Mining (EMM). The proposed framework identifies cohesive subgroups within data where model performance deviates exceptionally, highlighting regions of both high confidence and high uncertainty. We develop a new model class, mSMoPE (multiplex Soft Model Performance Evaluation), which quantifies uncertainty through conformal prediction's rigorous coverage guarantees. By defining a new quality measure, Relative Average Uncertainty Loss (RAUL), our framework isolates subgroups with exceptional performance patterns in multi-class classification and regression tasks. Experimental results across diverse datasets demonstrate the framework's effectiveness in uncovering interpretable subgroups that provide critical insights into model behavior. This work lays the groundwork for enhancing model interpretability and reliability, advancing the state-of-the-art in explainable AI and uncertainty quantification.