Analysis and Evaluation of Synthetic Data Generation in Speech Dysfluency Detection

TL;DR

This paper introduces LLM-Dys, a comprehensive synthetic speech dataset with enhanced dysfluency simulation, improving dysfluency detection accuracy and addressing data scarcity issues in clinical speech analysis.

Contribution

The paper presents LLM-Dys, a large-scale dysfluency speech corpus with LLM-based simulation, and an improved detection framework achieving state-of-the-art results.

Findings

State-of-the-art dysfluency detection performance achieved

LLM-Dys dataset covers 11 dysfluency categories

Open-source release of data, models, and code

Abstract

Speech dysfluency detection is crucial for clinical diagnosis and language assessment, but existing methods are limited by the scarcity of high-quality annotated data. Although recent advances in TTS model have enabled synthetic dysfluency generation, existing synthetic datasets suffer from unnatural prosody and limited contextual diversity. To address these limitations, we propose LLM-Dys -- the most comprehensive dysfluent speech corpus with LLM-enhanced dysfluency simulation. This dataset captures 11 dysfluency categories spanning both word and phoneme levels. Building upon this resource, we improve an end-to-end dysfluency detection framework. Experimental validation demonstrates state-of-the-art performance. All data, models, and code are open-sourced at https://github.com/Berkeley-Speech-Group/LLM-Dys.