Three new papers published

We have published three new papers (currently under peer review) on ArXiv, covering speaker anonymization, audio watermarking, and language model bias. If you are interested, please click on each paper to view the preprints of each paper.

Natalia Tomashenko, Junichi Yamagishi, Xin Wang, Yun Liu, Emmanuel Vincent, “Target speaker anonymization in multi-speaker recordings,” Oct. 2025

Preprint: https://arxiv.org/abs/2510.09307
Audio sample: https://sites.google.com/view/target-speaker-anonymization

Wanying Ge, Xin Wang, Junichi Yamagishi, “FakeMark: Deepfake Speech Attribution With Watermarked Artifacts“, Sept. 2025

Preprint: https://arxiv.org/abs/2510.12042
Audio sample: https://fakemark-demo.github.io/fakemark-demo/

Guoqing Luo, Iffat Maab, Lili Mou, Junichi Yamagishi, “ Investigating Thinking Behaviors of Reasoning-Based Language Models for Social Bias Mitigation ”, Sept. 2025 (in collaboration with the University of Alberta, Canada)

Preprint: https://arxiv.org/abs/2510.17062

Back to list