Automatically Balancing Model Accuracy and Complexity using Solution and Fitness Evolution (SAFE)

TL;DR

This paper presents SAFE, a coevolutionary algorithm that dynamically balances model accuracy and complexity, achieving comparable performance to standard methods on complex genetic datasets.

Contribution

The paper introduces SAFE, a novel coevolutionary algorithm that automatically tunes multiple objectives without performance loss in biomedical data modeling.

Findings

SAFE effectively balances accuracy and complexity.

SAFE matches standard algorithms' performance on genetic datasets.

SAFE operates without performance degradation.

Abstract

When seeking a predictive model in biomedical data, one often has more than a single objective in mind, e.g., attaining both high accuracy and low complexity (to promote interpretability). We investigate herein whether multiple objectives can be dynamically tuned by our recently proposed coevolutionary algorithm, SAFE (Solution And Fitness Evolution). We find that SAFE is able to automatically tune accuracy and complexity with no performance loss, as compared with a standard evolutionary algorithm, over complex simulated genetics datasets produced by the GAMETES tool.