Generalized Source Tracing: Detecting Novel Audio Deepfake Algorithm with Real Emphasis and Fake Dispersion Strategy

TL;DR

This paper introduces the REFD strategy for detecting and attributing audio deepfake algorithms, effectively distinguishing known and novel deepfake sources with high accuracy, addressing challenges posed by rapidly evolving deepfake techniques.

Contribution

The paper proposes the REFD strategy and a novel OOD method NSD for improved detection and attribution of both known and new deepfake audio algorithms.

Findings

REFD achieves 86.83% F1-score in Audio Deepfake Detection Challenge 2023

NSD effectively identifies novel deepfake algorithms using feature and logits similarity

The approach outperforms existing methods in OOD detection for deepfake audio

Abstract

With the proliferation of deepfake audio, there is an urgent need to investigate their attribution. Current source tracing methods can effectively distinguish in-distribution (ID) categories. However, the rapid evolution of deepfake algorithms poses a critical challenge in the accurate identification of out-of-distribution (OOD) novel deepfake algorithms. In this paper, we propose Real Emphasis and Fake Dispersion (REFD) strategy for audio deepfake algorithm recognition, demonstrating its effectiveness in discriminating ID samples while identifying OOD samples. For effective OOD detection, we first explore current post-hoc OOD methods and propose NSD, a novel OOD approach in identifying novel deepfake algorithms through the similarity consideration of both feature and logits scores. REFD achieves 86.83% F1-score as a single system in Audio Deepfake Detection Challenge 2023 Track3, showcasing its state-of-the-art performance.