Leveraging ASR Pretrained Conformers for Speaker Verification through Transfer Learning and Knowledge Distillation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cai, Danwei, Li, Ming
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913431608623104
author Cai, Danwei
Li, Ming
author_facet Cai, Danwei
Li, Ming
contents This paper explores the use of ASR-pretrained Conformers for speaker verification, leveraging their strengths in modeling speech signals. We introduce three strategies: (1) Transfer learning to initialize the speaker embedding network, improving generalization and reducing overfitting. (2) Knowledge distillation to train a more flexible speaker verification model, incorporating frame-level ASR loss as an auxiliary task. (3) A lightweight speaker adaptor for efficient feature conversion without altering the original ASR Conformer, allowing parallel ASR and speaker verification. Experiments on VoxCeleb show significant improvements: transfer learning yields a 0.48% EER, knowledge distillation results in a 0.43% EER, and the speaker adaptor approach, with just an added 4.92M parameters to a 130.94M-parameter model, achieves a 0.57% EER. Overall, our methods effectively transfer ASR capabilities to speaker verification tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2309_03019
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Leveraging ASR Pretrained Conformers for Speaker Verification through Transfer Learning and Knowledge Distillation
Cai, Danwei
Li, Ming
Audio and Speech Processing
This paper explores the use of ASR-pretrained Conformers for speaker verification, leveraging their strengths in modeling speech signals. We introduce three strategies: (1) Transfer learning to initialize the speaker embedding network, improving generalization and reducing overfitting. (2) Knowledge distillation to train a more flexible speaker verification model, incorporating frame-level ASR loss as an auxiliary task. (3) A lightweight speaker adaptor for efficient feature conversion without altering the original ASR Conformer, allowing parallel ASR and speaker verification. Experiments on VoxCeleb show significant improvements: transfer learning yields a 0.48% EER, knowledge distillation results in a 0.43% EER, and the speaker adaptor approach, with just an added 4.92M parameters to a 130.94M-parameter model, achieves a 0.57% EER. Overall, our methods effectively transfer ASR capabilities to speaker verification tasks.
title Leveraging ASR Pretrained Conformers for Speaker Verification through Transfer Learning and Knowledge Distillation
topic Audio and Speech Processing
url https://arxiv.org/abs/2309.03019