Beyond Hard Sharing: Efficient Multi-Task Speech-to-Text Modeling with Supervised Mixture of Experts

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Jin, Hojun, Hong, Eunsoo, Hyung, Ziwon, Lim, Sungjun, Lee, Seungjin, Cho, Keunseok
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916897923006464
author Jin, Hojun
Hong, Eunsoo
Hyung, Ziwon
Lim, Sungjun
Lee, Seungjin
Cho, Keunseok
author_facet Jin, Hojun
Hong, Eunsoo
Hyung, Ziwon
Lim, Sungjun
Lee, Seungjin
Cho, Keunseok
contents Hard-parameter sharing is a common strategy to train a single model jointly across diverse tasks. However, this often leads to task interference, impeding overall model performance. To address the issue, we propose a simple yet effective Supervised Mixture of Experts (S-MoE). Unlike traditional Mixture of Experts models, S-MoE eliminates the need for training gating functions by utilizing special guiding tokens to route each task to its designated expert. By assigning each task to a separate feedforward network, S-MoE overcomes the limitations of hard-parameter sharing. We further apply S-MoE to a speech-to-text model, enabling the model to process mixed-bandwidth input while jointly performing automatic speech recognition (ASR) and speech translation (ST). Experimental results demonstrate the effectiveness of the proposed S-MoE, achieving a 6.35% relative improvement in Word Error Rate (WER) when applied to both the encoder and decoder.
format Preprint
id arxiv_https___arxiv_org_abs_2508_10009
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond Hard Sharing: Efficient Multi-Task Speech-to-Text Modeling with Supervised Mixture of Experts
Jin, Hojun
Hong, Eunsoo
Hyung, Ziwon
Lim, Sungjun
Lee, Seungjin
Cho, Keunseok
Computation and Language
Artificial Intelligence
Sound
Audio and Speech Processing
Hard-parameter sharing is a common strategy to train a single model jointly across diverse tasks. However, this often leads to task interference, impeding overall model performance. To address the issue, we propose a simple yet effective Supervised Mixture of Experts (S-MoE). Unlike traditional Mixture of Experts models, S-MoE eliminates the need for training gating functions by utilizing special guiding tokens to route each task to its designated expert. By assigning each task to a separate feedforward network, S-MoE overcomes the limitations of hard-parameter sharing. We further apply S-MoE to a speech-to-text model, enabling the model to process mixed-bandwidth input while jointly performing automatic speech recognition (ASR) and speech translation (ST). Experimental results demonstrate the effectiveness of the proposed S-MoE, achieving a 6.35% relative improvement in Word Error Rate (WER) when applied to both the encoder and decoder.
title Beyond Hard Sharing: Efficient Multi-Task Speech-to-Text Modeling with Supervised Mixture of Experts
topic Computation and Language
Artificial Intelligence
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2508.10009