New papers tackle audio deepfake detection robustness

Arxiv CS.SD (Sound/Audio ML)

Multiple new preprints on arXiv propose methods for detecting synthetic speech, environmental sounds, and AI-generated songs, focusing on generalization across unseen generators and domain adaptation. Techniques include temporal pyramid modeling and phonetic explainability.

Deepfake detectionSpeech spoofing

See 9 more sources

Beyond Artifacts: Towards Generalizable Synthetic Song Detection via Music-Intrinsic Features5 days agoYan Han, Zhibin Wen, Yuan Wang, Shuangrun Shao, Xiaobing Li, Yang Xu, Wei Li

A Multi-Domain Feature Fusion Framework for Generalizable Deepfake Detection Across Different Generators7 days agoAmna Amjid, Sana Qadir, Mehwish Fatima, Raja Khurram Shahzad

GRIDEX: Grid-Grounded Forensic Explanations for Deepfake Spectrogram Analysis4 days agoThi Ngan Ha Do, Tingmin Wu, Alsharif Abuadbba, Kristen Moore

From Self-Supervised Speech Models to Mixture-of-Experts for Robust Anti-Spoofing7 days agoHugo Daumain, Driss Matrouf, Khaled Khelif, Mickael Rouvier

Teacher-Student Structure for Domain Adaptation in Ensemble Audio-Visual Video Deepfake Detection5 days agoElham Abolhasani, Maryam Ramezani, Hamid R. Rabiee

Phonetically Explainable Speech Deepfake Detection5 days agoManasi Chhibber, Jagabandhu Mishra, Tomi H. Kinnunen

Robust Spoofed Speech Detection via Temporal Pyramid Modeling5 days agoMahtab Masoudi Nezhad, Nima Karimian

HyperPotter: Spell the Charm of High-Order Interactions in Audio Deepfake Detection7 days agoQing Wen, Haohao Li, Zhongjie Ba, Peng Cheng, Miao He, Li Lu, Kui Ren

Dual-Granularity Orthogonal Disentanglement for Generalizable Audio Deepfake Detection5 days agoZhuodong Liu, Hugen Lv, Xiangyu Li, Chunhong Yuan

Arxiv CS.SD (Sound/Audio ML)·Han YinYang XiaoRohan Kumar DasJisheng BaiTing Dang

Jun 18, 4:00 AM