Adaptive Compression and Targeted Perturbation: A Unified Framework for Generalized Audio Deepfake Detection

Existing audio deepfake detection ADD methods frequently exhibit limited generalization, as fixed bottleneck architectures fail to adapt to the heterogeneous information densities inherent in diverse spoofing artifacts. Moreover, standard augmentation often leads to a pathological...

Detection-Guided Adaptive Purification with Diffusion Models for Robust Audio Deepfake Detection

Audio deepfake detectors remain vulnerable to adversarial perturbations that suppress the acoustic cues used for detection, allowing manipulated utterances to evade the detector. Although existing defenses can improve robustness, they require retraining the detector or introduce additional...

On Evaluating Quantum Kernel Robustness for Low-Resource Cross-Corpus Audio Deepfake Detection

Synthetic speech detection is critical for audio security, but performance can degrade when labeled data are scarce and evaluation conditions differ from training. This study examines quantum kernel methods and lightweight neural models for cross-corpus audio deepfake detection under limited...

Robust Workflow Generation Via Adversarial Learning for Audio Deepfake Detection

The rapid advancement of speech synthesis and voice conversion technologies has made audio deepfakes increasingly realistic, posing serious security risks in practical applications. While existing detection methods achieve strong performance under controlled conditions, they often fail to...

Rehearsal with Auxiliary-Informed Sampling for Audio Deepfake Detection

The performance of existing audio deepfake detection frameworks degrades when confronted with new deepfake attacks. Rehearsal-based continual learning CL, which updates models using a limited set of old data samples, helps preserve prior knowledge while incorporating new information. However,...

Top 5 AI-Powered Social Engineering Attacks

Social engineering has long been an effective tactic because of how it focuses on human vulnerabilities. There's no brute-force 'spray and pray' password guessing. No scouring systems for unpatched software. Instead, it simply relies on manipulating emotions such as trust, fear, and respect for...