# Towards Hierarchical Spoken Language Dysfluency Modeling

Source: [Towards Hierarchical Spoken Language Dysfluency Modeling](<https://www.sota2.com/research/paper/68fa71f51e209131cea53c5e>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T02:15:27.440Z

Authors: Jiachen Lian, Gopala Anumanchipalli

Published: 2024-01-18T14:33:01.000Z

## Abstract

Speech disfluency modeling is the bottleneck for both speech therapy and language learning. However, there is no effective AI solution to systematically tackle this problem. We solidify the concept of disfluent speech and disfluent speech modeling. We then present Hierarchical Unconstrained Disfluency Modeling (H-UDM) approach, the hierarchical extension of UDM that addresses both disfluency transcription and detection to eliminate the need for extensive manual annotation. Our experimental findings serve as clear evidence of the effectiveness and reliability of the methods we have introduced, encompassing both transcription and detection tasks.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Phonetic Transcription on Libri-Dys (test)](<https://www.sota2.com/research/sota/phonetic-transcription-on-libri-dys-test.md>): Phonetic Transcription; Libri-Dys (test); F1 Score: 88.9.

- [Phonetic Transcription on VCTK++ (test)](<https://www.sota2.com/research/sota/phonetic-transcription-on-vctk-plus-plus-test.md>): Phonetic Transcription; VCTK++ (test); F1 Score: 91.3.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2401.10015>)

[PDF](<https://arxiv.org/pdf/2401.10015.pdf>)