Mixture of Phonetic Experts Based Low-Rank Adaptation of Conformer Models for Accented English Speech Recognition

Not provided in the abstract

fine-tuningspeech recognitionphonetics

2610.26888

Builder Relevance

80%

2h ago

Abstract

A parameter-efficient adaptation framework for accented English speech recognition is introduced, focusing on phonetic categories rather than accent identity.

Reality Card

Core Claim

MoPE-LoRA achieves a 12.3% relative improvement in zero-shot evaluation over single LoRA by using phonetic-category-based decomposition.

Method / Result

12.3% relative improvement in zero-shot evaluation.

Limitations

The paper does not specify the authors, which may limit reproducibility and understanding of the context.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.

← Back to all papers