Introduces the LinguAS dataset of genuine and deepfaked audio annotated with expert-defined linguistic features to improve detection model performance over ASVspoof 2021 and SSL baselines.
Asvspoof 2021: accelerating progress in spoofed and deepfake speech detection
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
representative citing papers
EchoFake is a new replay-aware dataset combining zero-shot TTS deepfakes and physical replay recordings to improve generalization of speech deepfake detection models over existing lab-focused datasets.
Alethia is a pretrained audio encoder using continuous embedding prediction and generative flow-matching reconstruction that outperforms existing speech foundation models on voice deepfake tasks with better robustness and zero-shot generalization.
MLAAD provides a large-scale multi-language synthetic audio dataset for training and evaluating audio anti-spoofing models, showing better training performance than InTheWild and FakeOrReal and alternating superiority with ASVspoof 2019 across eight test sets.
Fairness metrics uncover gender disparities in audio deepfake detection error distributions that standard Equal Error Rate metrics obscure.
Truncated SSL backbone with logistic classifier detects audio deepfakes on-device, claimed to outperform AASIST by 10% while running 40% faster, packaged as a browser plugin.
SpAArSIST sparsifies AASIST by swapping learned pooling for explicit magnitude-based scoring and mean aggregation, cutting compute 20.7% and improving In-the-Wild EER to 2.82%.
citing papers explorer
-
Linguistically Augmented Audio Speech Data (LinguAS)
Introduces the LinguAS dataset of genuine and deepfaked audio annotated with expert-defined linguistic features to improve detection model performance over ASVspoof 2021 and SSL baselines.
-
EchoFake: A Replay-Aware Dataset for Practical Speech Deepfake Detection
EchoFake is a new replay-aware dataset combining zero-shot TTS deepfakes and physical replay recordings to improve generalization of speech deepfake detection models over existing lab-focused datasets.
-
Alethia: A Foundational Encoder for Voice Deepfakes
Alethia is a pretrained audio encoder using continuous embedding prediction and generative flow-matching reconstruction that outperforms existing speech foundation models on voice deepfake tasks with better robustness and zero-shot generalization.
-
MLAAD: The Multi-Language Audio Anti-Spoofing Dataset
MLAAD provides a large-scale multi-language synthetic audio dataset for training and evaluating audio anti-spoofing models, showing better training performance than InTheWild and FakeOrReal and alternating superiority with ASVspoof 2019 across eight test sets.
-
Gender Fairness in Audio Deepfake Detection: Performance and Disparity Analysis
Fairness metrics uncover gender disparities in audio deepfake detection error distributions that standard Equal Error Rate metrics obscure.
-
Detecting Audio Deepfakes on the Edge:Lightweight SSL-Based Detection in a Browser Plugin
Truncated SSL backbone with logistic classifier detects audio deepfakes on-device, claimed to outperform AASIST by 10% while running 40% faster, packaged as a browser plugin.
-
SpAArSIST: Sparsified AASIST for Efficient and Reliable Anti-Spoofing
SpAArSIST sparsifies AASIST by swapping learned pooling for explicit magnitude-based scoring and mean aggregation, cutting compute 20.7% and improving In-the-Wild EER to 2.82%.
- The Watermark Shortcut: How Provenance Marking Sabotages Audio Deepfake Detection