Boundary-Aware Adversarial Filtering for Reliable Diagnosis under Extreme Class Imbalance

TL;DR

This paper introduces AF-SMOTE, a novel augmentation framework that synthesizes and filters minority class data to improve recall and calibration in highly imbalanced classification tasks, especially in healthcare.

Contribution

AF-SMOTE is a new mathematically motivated augmentation method that enhances minority class detection by filtering synthesized points with an adversarial discriminator and boundary utility model.

Findings

Outperforms strong oversampling baselines in recall and precision.

Yields the best calibration among tested methods.

Proven to improve F_beta surrogate without inflating Brier score.

Abstract

We study classification under extreme class imbalance where recall and calibration are both critical, for example in medical diagnosis scenarios. We propose AF-SMOTE, a mathematically motivated augmentation framework that first synthesizes minority points and then filters them by an adversarial discriminator and a boundary utility model. We prove that, under mild assumptions on the decision boundary smoothness and class-conditional densities, our filtering step monotonically improves a surrogate of F_beta (for beta >= 1) while not inflating Brier score. On MIMIC-IV proxy label prediction and canonical fraud detection benchmarks, AF-SMOTE attains higher recall and average precision than strong oversampling baselines (SMOTE, ADASYN, Borderline-SMOTE, SVM-SMOTE), and yields the best calibration. We further validate these gains across multiple additional datasets beyond MIMIC-IV. Our successful application of AF-SMOTE to a healthcare dataset using a proxy label demonstrates in a disease-agnostic way its practical value in clinical situations, where missing true positive cases in rare diseases can have severe consequences.