Speech Intelligibility Assessment of Dysarthric Speech by using Goodness of Pronunciation with Uncertainty Quantification

Eun Jung Yeo, Kwanghee Choi, Sunhee Kim, Minhwa Chung
5/28/2023
cs.SDcs.LGeess.AS

No official code found. Best implementation: juice500ml/dysarthria-gop (28 stars, MIT, last commit 2025-03-13).

Abstract

This paper proposes an improved Goodness of Pronunciation (GoP) that utilizes Uncertainty Quantification (UQ) for automatic speech intelligibility assessment for dysarthric speech. Current GoP methods rely heavily on neural network-driven overconfident predictions, which is unsuitable for assessing dysarthric speech due to its significant acoustic differences from healthy speech. To alleviate the problem, UQ techniques were used on GoP by 1) normalizing the phoneme prediction (entropy, margin, maxlogit, logit-margin) and 2) modifying the scoring function (scaling, prior normalization). As a result, prior-normalized maxlogit GoP achieves the best performance, with a relative increase of 5.66%, 3.91%, and 23.65% compared to the baseline GoP for English, Korean, and Tamil, respectively. Furthermore, phoneme analysis is conducted to identify which phoneme scores significantly correlate with intelligibility scores in each language.

View on arXivView PDF

Code Implementations(1)

juice500ml/dysarthria-gopCommunity100%
MIT

Official implementation of the paper "Speech Intelligibility Assessment of Dysarthric Speech by using Goodness of Pronunciation with Uncertainty Quantification" (Interspeech 2023)

285Shell, PythonFeb 19, 20231 years agoMIT

Similar Papers

Applications of Artificial Intelligence for Cross-language Intelligibility Assessment of Dysarthric Speech

85% match2025

A study on the impact of Self-Supervised Learning on automatic dysarthric speech assessment

84% match2023

Bridging ASR and LLMs for Dysarthric Speech Recognition: Benchmarking Self-Supervised and Generative Approaches

83% match2025

Voice Biomarker Analysis and Automated Severity Classification of Dysarthric Speech in a Multilingual Context

82% match2024

Bridging the Perceptual-Statistical Gap in Dysarthria Assessment: Why Machine Learning Still Falls Short

82% match2025

Data-Efficient ASR Personalization for Non-Normative Speech Using an Uncertainty-Based Phoneme Difficulty Score for Guided Sampling

82% match2025

Cite this paper

@article{yeo2023speech,
  title  = {Speech Intelligibility Assessment of Dysarthric Speech by using Goodness of Pronunciation with Uncertainty Quantification},
  author = {Eun Jung Yeo and Kwanghee Choi and Sunhee Kim and Minhwa Chung},
  year   = {2023},
  eprint = {2305.18392},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.18392v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.18392.svg)](https://codeofpaper.com/paper/2305.18392)

Discussion