Dynamic Multi-Expert Projectors with Stabilized Routing for Multilingual Speech Recognition

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Pandey, Isha, Mittal, Ashish, Bahuguna, Vartul, Ramakrishnan, Ganesh
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918308213686272
author Pandey, Isha
Mittal, Ashish
Bahuguna, Vartul
Ramakrishnan, Ganesh
author_facet Pandey, Isha
Mittal, Ashish
Bahuguna, Vartul
Ramakrishnan, Ganesh
contents Recent advances in LLM-based ASR connect frozen speech encoders with Large Language Models (LLMs) via lightweight projectors. While effective in monolingual settings, a single projector struggles to capture the diverse acoustic-to-semantic mappings required for multilingual ASR. To address this, we propose SMEAR-MoE, a stabilized Mixture-of-Experts projector that ensures dense gradient flow to all experts, preventing expert collapse while enabling cross-lingual sharing. We systematically compare monolithic, static multi-projector, and dynamic MoE designs across four Indic languages (Hindi, Marathi, Tamil, Telugu). Our SMEAR-MoE achieves strong performance, delivering upto a 7.6% relative WER reduction over the single-projector baseline, while maintaining comparable runtime efficiency. Analysis of expert routing further shows linguistically meaningful specialization, with related languages sharing experts. These results demonstrate that stable multi-expert projectors are key to scalable and robust multilingual ASR.
format Preprint
id arxiv_https___arxiv_org_abs_2601_19451
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Dynamic Multi-Expert Projectors with Stabilized Routing for Multilingual Speech Recognition
Pandey, Isha
Mittal, Ashish
Bahuguna, Vartul
Ramakrishnan, Ganesh
Computation and Language
Recent advances in LLM-based ASR connect frozen speech encoders with Large Language Models (LLMs) via lightweight projectors. While effective in monolingual settings, a single projector struggles to capture the diverse acoustic-to-semantic mappings required for multilingual ASR. To address this, we propose SMEAR-MoE, a stabilized Mixture-of-Experts projector that ensures dense gradient flow to all experts, preventing expert collapse while enabling cross-lingual sharing. We systematically compare monolithic, static multi-projector, and dynamic MoE designs across four Indic languages (Hindi, Marathi, Tamil, Telugu). Our SMEAR-MoE achieves strong performance, delivering upto a 7.6% relative WER reduction over the single-projector baseline, while maintaining comparable runtime efficiency. Analysis of expert routing further shows linguistically meaningful specialization, with related languages sharing experts. These results demonstrate that stable multi-expert projectors are key to scalable and robust multilingual ASR.
title Dynamic Multi-Expert Projectors with Stabilized Routing for Multilingual Speech Recognition
topic Computation and Language
url https://arxiv.org/abs/2601.19451