NeurIPS 2023说话人识别最新工作:语音解耦与自监督!由新加坡国家科技局、新国立、港理工和港中文(深圳)联合推出。

论文链接:
https://arxiv.org/abs/2310.01128





另外,附录中增加的实验验证了该项工作中提出的Gt对于动态信息建模的有效性和必要性。

更多实验结果和详细的实验设定欢迎查阅原文。作者团队期待更多的交流,指导和未来的合作机会!
[1] Anthony Larcher, Kong Aik Lee, Bin Ma, and Haizhou Li. Text-dependent speaker verification:
Classifiers, databases and rsr2015. Speech Communication, 60:56–77, 2014.
[2] Tianyan Zhou, Yong Zhao, Jinyu Li, Yifan Gong, and Jian Wu. CNN with phonetic attention
for text-independent speaker verification. In Proc. IEEE ASRU, pages 718–725, 2019.
[3] Tianchi Liu, Rohan Kumar Das, Kong Aik Lee, and Haizhou Li. Neural acoustic-phonetic
approach for speaker verification with phonetic attention mask. IEEE Signal Processing
Letters, 29:782–786, 2022.
[4] Siqi Zheng, Yun Lei, and Hongbin Suo. Phonetically-aware coupled network for short duration
text-independent speaker verification. In Proc. Interspeech, pages 926–930, 2020.
[5] Joon Son Chung, Arsha Nagrani, and Andrew Zisserman. VoxCeleb2: Deep speaker recognition. In Proc. Interspeech, pages 1086–1090, 2018.
[6] Kong Aik Lee, Qiongqiong Wang, and Takafumi Koshinaka. Xi-vector embedding for speaker
recognition. IEEE Signal Processing Letters, 28:1385–1389, 2021.
[7] Mitchell McLaren, Luciana Ferrer, Diego Castan, and Aaron Lawson. The speakers in the
wild (SITW) speaker recognition database. In Proc. Interspeech, pages 818–822, 2016.
