Guardado en:
Detalles Bibliográficos
Autores principales: Male, Prabash Reddy, Ray, Swayambhu Nath, Arsikere, Harish, Jaiswal, Akshat, Swarup, Prakhar, Sen, Prantik, Chakrabarty, Debmalya, Girish, K V Vijay, Bhave, Nikhil, Weber, Frederick, Bhattacharya, Sambuddha, Garimella, Sri
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2505.19774
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910968672419840
author Male, Prabash Reddy
Ray, Swayambhu Nath
Arsikere, Harish
Jaiswal, Akshat
Swarup, Prakhar
Sen, Prantik
Chakrabarty, Debmalya
Girish, K V Vijay
Bhave, Nikhil
Weber, Frederick
Bhattacharya, Sambuddha
Garimella, Sri
author_facet Male, Prabash Reddy
Ray, Swayambhu Nath
Arsikere, Harish
Jaiswal, Akshat
Swarup, Prakhar
Sen, Prantik
Chakrabarty, Debmalya
Girish, K V Vijay
Bhave, Nikhil
Weber, Frederick
Bhattacharya, Sambuddha
Garimella, Sri
contents Recent advancements in speech encoders have drawn attention due to their integration with Large Language Models for various speech tasks. While most research has focused on either causal or full-context speech encoders, there's limited exploration to effectively handle both streaming and non-streaming applications, while achieving state-of-the-art performance. We introduce DuRep, a Dual-mode Speech Representation learning setup, which enables a single speech encoder to function efficiently in both offline and online modes without additional parameters or mode-specific adjustments, across downstream tasks. DuRep-200M, our 200M parameter dual-mode encoder, achieves 12% and 11.6% improvements in streaming and non-streaming modes, over baseline encoders on Multilingual ASR. Scaling this approach to 2B parameters, DuRep-2B sets new performance benchmarks across ASR and non-ASR tasks. Our analysis reveals interesting trade-offs between acoustic and semantic information across encoder layers.
format Preprint
id arxiv_https___arxiv_org_abs_2505_19774
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DuRep: Dual-Mode Speech Representation Learning via ASR-Aware Distillation
Male, Prabash Reddy
Ray, Swayambhu Nath
Arsikere, Harish
Jaiswal, Akshat
Swarup, Prakhar
Sen, Prantik
Chakrabarty, Debmalya
Girish, K V Vijay
Bhave, Nikhil
Weber, Frederick
Bhattacharya, Sambuddha
Garimella, Sri
Audio and Speech Processing
Recent advancements in speech encoders have drawn attention due to their integration with Large Language Models for various speech tasks. While most research has focused on either causal or full-context speech encoders, there's limited exploration to effectively handle both streaming and non-streaming applications, while achieving state-of-the-art performance. We introduce DuRep, a Dual-mode Speech Representation learning setup, which enables a single speech encoder to function efficiently in both offline and online modes without additional parameters or mode-specific adjustments, across downstream tasks. DuRep-200M, our 200M parameter dual-mode encoder, achieves 12% and 11.6% improvements in streaming and non-streaming modes, over baseline encoders on Multilingual ASR. Scaling this approach to 2B parameters, DuRep-2B sets new performance benchmarks across ASR and non-ASR tasks. Our analysis reveals interesting trade-offs between acoustic and semantic information across encoder layers.
title DuRep: Dual-Mode Speech Representation Learning via ASR-Aware Distillation
topic Audio and Speech Processing
url https://arxiv.org/abs/2505.19774