The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech

Early detection is crucial for managing incurable disorders, particularly autism spectrum disorder (ASD). Unfortunately, a considerable number of individuals with ASD receive a late diagnosis or remain undiagnosed. Speech holds a critical role in ASD, as a significant number of affected individuals...

Full description

Saved in:
Bibliographic Details
Main Authors: Najla D. Al Futaisi, Björn W. Schuller, Fabien Ringeval, Maja Pantic
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-08-01
Series:Frontiers in Digital Health
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fdgth.2025.1274675/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849736952065032192
author Najla D. Al Futaisi
Björn W. Schuller
Björn W. Schuller
Fabien Ringeval
Maja Pantic
author_facet Najla D. Al Futaisi
Björn W. Schuller
Björn W. Schuller
Fabien Ringeval
Maja Pantic
author_sort Najla D. Al Futaisi
collection DOAJ
description Early detection is crucial for managing incurable disorders, particularly autism spectrum disorder (ASD). Unfortunately, a considerable number of individuals with ASD receive a late diagnosis or remain undiagnosed. Speech holds a critical role in ASD, as a significant number of affected individuals experience speech impairments or remain non-verbal. To address this, we use speech analysis for automatic ASD recognition in children by classifying their speech as either autistic or typically developing. However, due to the lack of large labelled datasets, we leverage two smaller datasets to explore deep transfer learning methods. We investigate two fine-tuning approaches: (1) Discriminative Fine-Tuning (D-FT), which is pre-trained on a related dataset before being tuned on a similar task, and (2) Wav2Vec 2.0 Fine-Tuning (W2V2-FT), which leverages self-supervised speech representations pre-trained on a larger, unrelated dataset. We perform two distinct classification tasks: (a) a binary task to determine typicality, classifying speech as either that of a typically developing (TD) child or an atypically developing (AD) child; and (b) a four-class diagnosis task, which further classifies atypical cases into ASD, dysphasia (DYS), or pervasive developmental disorder-not otherwise specified (NOS), alongside TD. This research aims to improve early recognition strategies, particularly for individuals with ASD. The findings suggest that transfer learning methods can be a valuable tool for autism recognition from speech. For the typicality classification task (TD vs. AD), the D-FT model achieved the highest test UAR (94.8%), outperforming W2V2-FT (91.5%). In the diagnosis task (TD, ASD, DYS, NOS), D-FT also demonstrated superior performance (60.9% UAR) compared to W2V2-FT (54.3%). These results highlight the potential of transfer learning for speech-based ASD recognition and underscore the challenges of multi-class classification with limited labeled data.
format Article
id doaj-art-535bbaf2e951459cbec96056985935af
institution DOAJ
issn 2673-253X
language English
publishDate 2025-08-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Digital Health
spelling doaj-art-535bbaf2e951459cbec96056985935af2025-08-20T03:07:06ZengFrontiers Media S.A.Frontiers in Digital Health2673-253X2025-08-01710.3389/fdgth.2025.12746751274675The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speechNajla D. Al Futaisi0Björn W. Schuller1Björn W. Schuller2Fabien Ringeval3Maja Pantic4GLAM – Group on Language, Audio & Music, Imperial College London, London, United KingdomGLAM – Group on Language, Audio & Music, Imperial College London, London, United KingdomChair of Embedded Intelligence for Health Care and Wellbeing, University of Augsburg, Augsburg, GermanyGrenoble INP, LIG, University Grenoble Alpes, Inria, CNRS, Grenoble, FranceiBUG – Intelligent Behaviour Understanding Group, Imperial College London, London, United KingdomEarly detection is crucial for managing incurable disorders, particularly autism spectrum disorder (ASD). Unfortunately, a considerable number of individuals with ASD receive a late diagnosis or remain undiagnosed. Speech holds a critical role in ASD, as a significant number of affected individuals experience speech impairments or remain non-verbal. To address this, we use speech analysis for automatic ASD recognition in children by classifying their speech as either autistic or typically developing. However, due to the lack of large labelled datasets, we leverage two smaller datasets to explore deep transfer learning methods. We investigate two fine-tuning approaches: (1) Discriminative Fine-Tuning (D-FT), which is pre-trained on a related dataset before being tuned on a similar task, and (2) Wav2Vec 2.0 Fine-Tuning (W2V2-FT), which leverages self-supervised speech representations pre-trained on a larger, unrelated dataset. We perform two distinct classification tasks: (a) a binary task to determine typicality, classifying speech as either that of a typically developing (TD) child or an atypically developing (AD) child; and (b) a four-class diagnosis task, which further classifies atypical cases into ASD, dysphasia (DYS), or pervasive developmental disorder-not otherwise specified (NOS), alongside TD. This research aims to improve early recognition strategies, particularly for individuals with ASD. The findings suggest that transfer learning methods can be a valuable tool for autism recognition from speech. For the typicality classification task (TD vs. AD), the D-FT model achieved the highest test UAR (94.8%), outperforming W2V2-FT (91.5%). In the diagnosis task (TD, ASD, DYS, NOS), D-FT also demonstrated superior performance (60.9% UAR) compared to W2V2-FT (54.3%). These results highlight the potential of transfer learning for speech-based ASD recognition and underscore the challenges of multi-class classification with limited labeled data.https://www.frontiersin.org/articles/10.3389/fdgth.2025.1274675/fullautism spectrum disorderchild speechartificial intelligencedeep learningtransformertransfer learning
spellingShingle Najla D. Al Futaisi
Björn W. Schuller
Björn W. Schuller
Fabien Ringeval
Maja Pantic
The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech
Frontiers in Digital Health
autism spectrum disorder
child speech
artificial intelligence
deep learning
transformer
transfer learning
title The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech
title_full The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech
title_fullStr The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech
title_full_unstemmed The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech
title_short The Noor Project: fair transformer transfer learning for autism spectrum disorder recognition from speech
title_sort noor project fair transformer transfer learning for autism spectrum disorder recognition from speech
topic autism spectrum disorder
child speech
artificial intelligence
deep learning
transformer
transfer learning
url https://www.frontiersin.org/articles/10.3389/fdgth.2025.1274675/full
work_keys_str_mv AT najladalfutaisi thenoorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT bjornwschuller thenoorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT bjornwschuller thenoorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT fabienringeval thenoorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT majapantic thenoorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT najladalfutaisi noorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT bjornwschuller noorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT bjornwschuller noorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT fabienringeval noorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech
AT majapantic noorprojectfairtransformertransferlearningforautismspectrumdisorderrecognitionfromspeech