Learning domain-heterogeneous speaker recognition systems with personalized continual federated learning

EURASIP Journal on Audio, Speech, and Music Processing

Methodology

Shugong Xu¹  Zhiyong Chen¹ 

[1] School of Communication and Information Engineering, Shanghai University, Shanghai, China;

关键词: Speaker recognition;    Federated learning;    Domain adaptation;    Continual learning;   

DOI  :  10.1186/s13636-023-00299-2

 received in 2023-04-22, accepted in 2023-08-10,  发布年份 2023

来源: Springer

【 摘 要 】

Speaker recognition, the process of automatically identifying a speaker based on individual characteristics in speech signals, presents significant challenges when addressing heterogeneous-domain conditions. Federated learning, a recent development in machine learning methods, has gained traction in privacy-sensitive tasks, such as personal voice assistants in home environments. However, its application in heterogeneous multi-domain scenarios for enhancing system customization remains underexplored. In this paper, we propose the utilization of federated learning in heterogeneous situations to enable adaptation across multiple domains. We also introduce a personalized federated learning algorithm designed to effectively leverage limited domain data, resulting in improved learning outcomes. Furthermore, we present a strategy for implementing the federated learning algorithm in practical, real-world continual learning scenarios, demonstrating promising results. The proposed federated learning method exhibits superior performance across a range of synthesized complex conditions and continual learning settings, compared to conventional training methods.

【 授权许可】

CC BY   
© Springer Nature Switzerland AG 2023

【 图 表 】

Fig. 4

13690_2023_1170_Article_IEq81.gif

Fig. 3

12888_2023_5172_Article_IEq5.gif

Fig. 7

13690_2023_1170_Article_IEq203.gif

13690_2023_1170_Article_IEq202.gif

13690_2023_1170_Article_IEq201.gif

12888_2023_5142_Article_IEq11.gif

12888_2023_5142_Article_IEq19.gif

Fig. 4

Fig. 1

Fig. 6

Fig. 5

Fig. 1

42004_2023_990_Article_IEq87.gif

40249_2023_1135_Article_IEq7.gif

13690_2023_1170_Article_IEq58.gif

Fig. 2

13690_2023_1170_Article_IEq162.gif

13690_2023_1170_Article_IEq53.gif

Fig. 6

Fig. 1

13690_2023_1170_Article_IEq159.gif

13690_2023_1170_Article_IEq49.gif

Fig. 2

13690_2023_1170_Article_IEq157.gif

13690_2023_1170_Article_IEq153.gif

13690_2023_1170_Article_IEq152.gif

13690_2023_1187_Article_IEq1.gif

Fig.1

Fig. 1

Fig. 9

Fig. 1

Fig. 1

Fig. 5

13063_2023_7612_Figbt_HTML.png

12888_2023_5172_Article_IEq41.gif

41408_2023_919_Article_IEq11.gif

12888_2023_5172_Article_IEq36.gif

Fig. 2

Fig. 5