In-context learning capabilities of Large Language Models to detect suicide risk among adolescents from speech transcripts

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Roquefort, Filomene, Ducorroy, Alexandre, Riad, Rachid
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918035158204416
author Roquefort, Filomene
Ducorroy, Alexandre
Riad, Rachid
author_facet Roquefort, Filomene
Ducorroy, Alexandre
Riad, Rachid
contents Early suicide risk detection in adolescents is critical yet hindered by scalability challenges of current assessments. This paper presents our approach to the first SpeechWellness Challenge (SW1), which aims to assess suicide risk in Chinese adolescents through speech analysis. Due to speech anonymization constraints, we focused on linguistic features, leveraging Large Language Models (LLMs) for transcript-based classification. Using DSPy for systematic prompt engineering, we developed a robust in-context learning approach that outperformed traditional fine-tuning on both linguistic and acoustic markers. Our systems achieved third and fourth places among 180+ submissions, with 0.68 accuracy (F1=0.7) using only transcripts. Ablation analyses showed that increasing prompt example improved performance (p=0.003), with varying effects across model types and sizes. These findings advance automated suicide risk assessment and demonstrate LLMs' value in mental health applications.
format Preprint
id arxiv_https___arxiv_org_abs_2505_20491
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle In-context learning capabilities of Large Language Models to detect suicide risk among adolescents from speech transcripts
Roquefort, Filomene
Ducorroy, Alexandre
Riad, Rachid
Audio and Speech Processing
Sound
Early suicide risk detection in adolescents is critical yet hindered by scalability challenges of current assessments. This paper presents our approach to the first SpeechWellness Challenge (SW1), which aims to assess suicide risk in Chinese adolescents through speech analysis. Due to speech anonymization constraints, we focused on linguistic features, leveraging Large Language Models (LLMs) for transcript-based classification. Using DSPy for systematic prompt engineering, we developed a robust in-context learning approach that outperformed traditional fine-tuning on both linguistic and acoustic markers. Our systems achieved third and fourth places among 180+ submissions, with 0.68 accuracy (F1=0.7) using only transcripts. Ablation analyses showed that increasing prompt example improved performance (p=0.003), with varying effects across model types and sizes. These findings advance automated suicide risk assessment and demonstrate LLMs' value in mental health applications.
title In-context learning capabilities of Large Language Models to detect suicide risk among adolescents from speech transcripts
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2505.20491