Towards Robust Audio Deepfake Detection: A Evolving Benchmark for Continual Learning

TL;DR

This paper introduces EVDA, a comprehensive benchmark for evaluating continual learning methods in deepfake audio detection, addressing the challenge of adapting to evolving synthetic speech while maintaining detection accuracy on older data.

Contribution

It presents a new benchmark, EVDA, that supports various continual learning techniques and datasets, fostering development of robust deepfake audio detection algorithms.

Findings

EVDA enables effective evaluation of continual learning methods.

Supports multiple datasets including GPT-4 generated audio.

Facilitates development of more robust deepfake detection algorithms.

Abstract

The rise of advanced large language models such as GPT-4, GPT-4o, and the Claude family has made fake audio detection increasingly challenging. Traditional fine-tuning methods struggle to keep pace with the evolving landscape of synthetic speech, necessitating continual learning approaches that can adapt to new audio while retaining the ability to detect older types. Continual learning, which acts as an effective tool for detecting newly emerged deepfake audio while maintaining performance on older types, lacks a well-constructed and user-friendly evaluation framework. To address this gap, we introduce EVDA, a benchmark for evaluating continual learning methods in deepfake audio detection. EVDA includes classic datasets from the Anti-Spoofing Voice series, Chinese fake audio detection series, and newly generated deepfake audio from models like GPT-4 and GPT-4o. It supports various continual learning techniques, such as Elastic Weight Consolidation (EWC), Learning without Forgetting (LwF), and recent methods like Regularized Adaptive Weight Modification (RAWM) and Radian Weight Modification (RWM). Additionally, EVDA facilitates the development of robust algorithms by providing an open interface for integrating new continual learning methods