FSIR: Few-Shot Speaker Identification using Reptile Algorithm

Speaker identification (SI) is the evolving biometrics and holds great potential for real-world applications. SI is challenging task with limited data in forensics. Few-shot learning inspires from human learning. In this work, SI tasks are investigated using optimization based meta-learning. A novel...

Full description

Saved in:
Bibliographic Details
Published in2023 8th International Conference on Computers and Devices for Communication (CODEC) pp. 1 - 2
Main Authors Saritha, Banala, K, Anish Monsley, Hussain Laskar, Rabul, Choudhury, Madhuchhanda
Format Conference Proceeding
LanguageEnglish
Published IEEE 14.12.2023
Subjects
Computer architecture
Convolutional Neural Networks
few-shot learning
Forensics
Metalearning
Performance evaluation
Recurrent neural networks
Reptile Algorithm
Speaker Identification
Statistical analysis
Online AccessGet full text

Cover

More Information
Summary:Speaker identification (SI) is the evolving biometrics and holds great potential for real-world applications. SI is challenging task with limited data in forensics. Few-shot learning inspires from human learning. In this work, SI tasks are investigated using optimization based meta-learning. A novel architecture for SI system is proposed and is based on a convolutional recurrent neural network. The accuracy is improved by 3% using the proposed framework, outperforming statistical methods.
DOI:10.1109/CODEC60112.2023.10466164