Explainable-by-Design Audio Deepfake Detection via Wiener-Hopf Linear Prediction

2026-07-14

AI Summary

This work addresses the growing challenge of detecting audio deepfakes amid the proliferation of synthetic speech by proposing an interpretable detection framework based on Wiener-Hopf linear prediction. The approach integrates classical signal processing theory with a lightweight 2D convolutional neural network through a “design-for-interpretability” architecture, which directly links linear prediction coefficients to classification decisions, thereby significantly enhancing model transparency. Evaluated on multiple standard datasets, the method achieves state-of-the-art detection accuracy while substantially reducing computational complexity. Furthermore, it demonstrates strong robustness against common post-processing operations, including additive noise, MP3 compression, and telephone-band filtering.

Abstract

The rapid advancement of synthetic speech generation methods has made audio deepfake detection a critical challenge in multimedia forensics. While recent approaches achieve high detection accuracy, they typically rely on black-box architectures that offer limited interpretability and high computational complexity. In this paper, we propose an explainable-by-design audio deepfake detection framework based on Wiener-Hopf linear prediction, processed by a lightweight 2D Convolutional Neural Network (CNN). This design enables a direct and transparent connection between classification outcomes and the acoustic properties of the signal. Experimental results on benchmark datasets demonstrate competitive detection performance while maintaining significantly lower computational complexity compared to state-of-the-art solutions. The interpretability analysis using Grad-CAM reveals that the classifier focuses on low-order predictor coefficients and on silence and transitional regions, suggesting that the Wiener-Hopf predictor captures reverberation characteristics and subtle statistical inconsistencies in synthetic speech. Finally, robustness experiments show that fine-tuning effectively recovers detection performance under common post-processing degradations, including additive noise, MP3 compression, and telephone filtering.

Problem

• audio deepfake detection
• explainability
• computational complexity
• multimedia forensics
• synthetic speech

Innovation

• explainable-by-design
• Wiener-Hopf linear prediction
• audio deepfake detection
• lightweight CNN
• interpretability