Targeted Augmented Data for Audio Deepfake Detection
This addresses the need for robust audio deepfake detectors to prevent overfitting and enhance performance against unseen manipulations, representing an incremental improvement in the field.
The paper tackles the problem of audio deepfake detection by proposing a novel augmentation method that generates pseudo-fakes to target the model's decision boundary, resulting in improved generalization capabilities as demonstrated in experiments on two architectures.
The availability of highly convincing audio deepfake generators highlights the need for designing robust audio deepfake detectors. Existing works often rely solely on real and fake data available in the training set, which may lead to overfitting, thereby reducing the robustness to unseen manipulations. To enhance the generalization capabilities of audio deepfake detectors, we propose a novel augmentation method for generating audio pseudo-fakes targeting the decision boundary of the model. Inspired by adversarial attacks, we perturb original real data to synthesize pseudo-fakes with ambiguous prediction probabilities. Comprehensive experiments on two well-known architectures demonstrate that the proposed augmentation contributes to improving the generalization capabilities of these architectures.