[INTERSPEECH '26] Does Fine-tuning by Reinforcement Learning Improve Generalization in Binary Speech Deepfake Detection?
Xin Wang, Wanying Ge, Junichi Yamagishi
in INTERSPEECH 2026.

[CVPR Workshop '26] FakeMark: Deepfake Speech Attribution With Watermarked Artifacts
Wanying Ge, Xin Wang, Junichi Yamagishi
The Authenticity & Provenance in the Age of AI workshop.
The local museum is one of the best I visited.

[ASRU '25] Post-training for Deepfake Speech Detection
Wanying Ge, Xin Wang, Xuechen Liu, Junichi Yamagishi
in Proc. IEEE ASRU 2025, pp. 1-8.
This song came to my mind every time downloading and preparing large databases.

[INTERSPEECH '25] A Comparative Study on Proactive and Passive Detection of Deepfake Speech
Chia-Hua Wu, Wanying Ge, Xin Wang, Junichi Yamagishi, Yu Tsao, Hsin-Min Wang
in Proc. INTERSPEECH 2025, pp. 5328-5332.

[Odyssey '22] Explainable Deepfake and Spoofing Detection: An Attack Analysis Using SHapley Additive exPlanations
Wanying Ge, Massimiliano Todisco, Nicholas Evans
in Proc. The Speaker and Language Recognition Workshop (Odyssey 2022), pp. 70-76.
The statement we wrote in Section 4, “We accept that, within our current framework, it is impossible to fully isolate classifier behaviour from attack artefacts, since the latter are always learned using a specific classifier.” was inspired by this book. A similar concept can be found in Chapters 2&3 of this book.

[ICASSP '22] Explaining Deep Learning Models for Spoofing and Deepfake Detection With SHapley Additive exPlanations
Wanying Ge, Jose Patino, Massimiliano Todisco, Nicholas Evans
in Proc. ICASSP 2022, pp. 6387-6391.
I was attending a session on my Master topic when I realized I could apply the same method to my PhD topic.

Monaural Noisy Speech Separation Combining Sparse Non-negative Matrix Factorization and Deep Attractor Network
Wanying Ge, Tianqi Zhang, Congcong Fan, Tian Zhang
Chinese Journal of Acoustics, 2020.
In my thesis, another version of this figure is a mixture of a female voice “I am the mask you wear” and a male voice “It’s me they hear”, from this song, since the task is to estimate a time-frequency mask.