Evolutionary Multi-Objective Fusion of Deepfake Speech Detectors

TL;DR

This paper introduces an evolutionary multi-objective fusion framework for deepfake speech detectors that balances detection accuracy with system complexity, achieving state-of-the-art results with fewer parameters.

Contribution

It proposes a novel NSGA-II based multi-objective score fusion method that optimizes both detection performance and system complexity for SSL-based deepfake speech detectors.

Findings

Pareto fronts outperform simple averaging and logistic regression baselines.

Real-valued variant achieves 2.37% EER and reduces parameters by half.

Provides diverse trade-off solutions for deployment balancing accuracy and complexity.

Abstract

While deepfake speech detectors built on large self-supervised learning (SSL) models achieve high accuracy, employing standard ensemble fusion to further enhance robustness often results in oversized systems with diminishing returns. To address this, we propose an evolutionary multi-objective score fusion framework that jointly minimizes detection error and system complexity. We explore two encodings optimized by NSGA-II: binary-coded detector selection for score averaging and a real-valued scheme that optimizes detector weights for a weighted sum. Experiments on the ASVspoof 5 dataset with 36 SSL-based detectors show that the obtained Pareto fronts outperform simple averaging and logistic regression baselines. The real-valued variant achieves 2.37% EER (0.0684 minDCF) and identifies configurations that match state-of-the-art performance while significantly reducing system complexity, requiring only half the parameters. Our method also provides a diverse set of trade-off solutions, enabling deployment choices that balance accuracy and computational cost.