Title: SHIELD: A Secure and Highly Enhanced Integrated Learning for Robust Deepfake Detection against Adversarial Attacks Show Chinese title

Title: SHIELD：一种针对对抗攻击的鲁棒深度伪造检测的安全且高度增强的集成学习方法

Abstract: Audio plays a crucial role in applications like speaker verification, voice-enabled smart devices, and audio conferencing. However, audio manipulations, such as deepfakes, pose significant risks by enabling the spread of misinformation. Our empirical analysis reveals that existing methods for detecting deepfake audio are often vulnerable to anti-forensic (AF) attacks, particularly those attacked using generative adversarial networks. In this article, we propose a novel collaborative learning method called SHIELD to defend against generative AF attacks. To expose AF signatures, we integrate an auxiliary generative model, called the defense (DF) generative model, which facilitates collaborative learning by combining input and output. Furthermore, we design a triplet model to capture correlations for real and AF attacked audios with real-generated and attacked-generated audios using auxiliary generative models. The proposed SHIELD strengthens the defense against generative AF attacks and achieves robust performance across various generative models. The proposed AF significantly reduces the average detection accuracy from 95.49% to 59.77% for ASVspoof2019, from 99.44% to 38.45% for In-the-Wild, and from 98.41% to 51.18% for HalfTruth for three different generative models. The proposed SHIELD mechanism is robust against AF attacks and achieves an average accuracy of 98.13%, 98.58%, and 99.57% in match, and 98.78%, 98.62%, and 98.85% in mismatch settings for the ASVspoof2019, In-the-Wild, and HalfTruth datasets, respectively. Show Chinese abstract

Abstract: 音频在语音验证、语音启用的智能设备和音频会议等应用中起着至关重要的作用。 然而，音频篡改，如深度伪造，通过传播虚假信息带来了重大风险。 我们的实证分析表明，现有的深度伪造音频检测方法通常容易受到反取证(AF)攻击，特别是那些使用生成对抗网络进行攻击的方法。 在本文中，我们提出了一种称为SHIELD的新协同学习方法，以抵御生成型AF攻击。 为了暴露AF特征，我们集成了一个辅助生成模型，称为防御(DF)生成模型，该模型通过结合输入和输出来促进协同学习。 此外，我们设计了一个三元组模型，利用辅助生成模型捕获真实音频和AF攻击音频与真实生成音频和攻击生成音频之间的相关性。 提出的SHIELD增强了对生成型AF攻击的防御，并在各种生成模型上实现了稳健的性能。 所提出的AF显著降低了ASVspoof2019的平均检测准确率，从95.49%降至59.77%，In-the-Wild的平均检测准确率从99.44%降至38.45%，HalfTruth的平均检测准确率从98.41%降至51.18%，针对三种不同的生成模型。 所提出的SHIELD机制对AF攻击具有鲁棒性，在ASVspoof2019、In-the-Wild和HalfTruth数据集的匹配设置中，分别实现了98.13%、98.58%和99.57%的平均准确率，在非匹配设置中分别实现了98.78%、98.62%和98.85%的平均准确率。