# Time and Tokens: Benchmarking End-to-End Speech Dysfluency Detection

Source: [Time and Tokens: Benchmarking End-to-End Speech Dysfluency Detection](<https://www.sota2.com/research/paper/68fa74e61e209131cea648f5>)

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

Data updated: 2025-10-26T03:21:26.418Z

Authors: Xuanru Zhou, Jiachen Lian, Cheol Jun Cho, Jingwen Liu, Zongli Ye, Jinming Zhang, Brittany Morin, David Baquirin, Jet Vonk, Zoe Ezzes, Zachary Miller, Maria Luisa Gorno Tempini, Gopala Anumanchipalli

Published: 2024-09-20T15:35:32.000Z

## Abstract

Speech dysfluency modeling is a task to detect dysfluencies in speech, such as repetition, block, insertion, replacement, and deletion. Most recent advancements treat this problem as a time-based object detection problem. In this work, we revisit this problem from a new perspective: tokenizing dysfluencies and modeling the detection problem as a token-based automatic speech recognition (ASR) problem. We propose rule-based speech and text dysfluency simulators and develop VCTK-token, and then develop a Whisper-like seq2seq architecture to build a new benchmark with decent performance. We also systematically compare our proposed token-based methods with time-based methods, and propose a unified benchmark to facilitate future research endeavors. We open-source these resources for the broader scientific community. The project page is available at https://rorizzz.github.io/

## 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.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2409.13582>)

[PDF](<https://arxiv.org/pdf/2409.13582.pdf>)
