HDMoLE: Mixture of LoRA Experts with Hierarchical Routing and Dynamic Thresholds for Fine-Tuning LLM-based ASR Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mu, Bingshen, Wei, Kun, Shao, Qijie, Xu, Yong, Xie, Lei
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916549252612096
author Mu, Bingshen
Wei, Kun
Shao, Qijie
Xu, Yong
Xie, Lei
author_facet Mu, Bingshen
Wei, Kun
Shao, Qijie
Xu, Yong
Xie, Lei
contents Recent advancements in integrating Large Language Models (LLM) with automatic speech recognition (ASR) have performed remarkably in general domains. While supervised fine-tuning (SFT) of all model parameters is often employed to adapt pre-trained LLM-based ASR models to specific domains, it imposes high computational costs and notably reduces their performance in general domains. In this paper, we propose a novel parameter-efficient multi-domain fine-tuning method for adapting pre-trained LLM-based ASR models to multi-accent domains without catastrophic forgetting named \textit{HDMoLE}, which leverages hierarchical routing and dynamic thresholds based on combining low-rank adaptation (LoRA) with the mixer of experts (MoE) and can be generalized to any linear layer. Hierarchical routing establishes a clear correspondence between LoRA experts and accent domains, improving cross-domain collaboration among the LoRA experts. Unlike the static Top-K strategy for activating LoRA experts, dynamic thresholds can adaptively activate varying numbers of LoRA experts at each MoE layer. Experiments on the multi-accent and standard Mandarin datasets demonstrate the efficacy of HDMoLE. Applying HDMoLE to an LLM-based ASR model projector module achieves similar performance to full fine-tuning in the target multi-accent domains while using only 9.6% of the trainable parameters required for full fine-tuning and minimal degradation in the source general domain.
format Preprint
id arxiv_https___arxiv_org_abs_2409_19878
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HDMoLE: Mixture of LoRA Experts with Hierarchical Routing and Dynamic Thresholds for Fine-Tuning LLM-based ASR Models
Mu, Bingshen
Wei, Kun
Shao, Qijie
Xu, Yong
Xie, Lei
Sound
Audio and Speech Processing
Recent advancements in integrating Large Language Models (LLM) with automatic speech recognition (ASR) have performed remarkably in general domains. While supervised fine-tuning (SFT) of all model parameters is often employed to adapt pre-trained LLM-based ASR models to specific domains, it imposes high computational costs and notably reduces their performance in general domains. In this paper, we propose a novel parameter-efficient multi-domain fine-tuning method for adapting pre-trained LLM-based ASR models to multi-accent domains without catastrophic forgetting named \textit{HDMoLE}, which leverages hierarchical routing and dynamic thresholds based on combining low-rank adaptation (LoRA) with the mixer of experts (MoE) and can be generalized to any linear layer. Hierarchical routing establishes a clear correspondence between LoRA experts and accent domains, improving cross-domain collaboration among the LoRA experts. Unlike the static Top-K strategy for activating LoRA experts, dynamic thresholds can adaptively activate varying numbers of LoRA experts at each MoE layer. Experiments on the multi-accent and standard Mandarin datasets demonstrate the efficacy of HDMoLE. Applying HDMoLE to an LLM-based ASR model projector module achieves similar performance to full fine-tuning in the target multi-accent domains while using only 9.6% of the trainable parameters required for full fine-tuning and minimal degradation in the source general domain.
title HDMoLE: Mixture of LoRA Experts with Hierarchical Routing and Dynamic Thresholds for Fine-Tuning LLM-based ASR Models
topic Sound
Audio and Speech Processing
url https://arxiv.org/abs/2409.19878