DAIC-WOZ: On the Validity of Using the Therapist's prompts in Automatic Depression Detection from Clinical Interviews

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F26%3A0196771" target="_blank" >RIV/00216305:26230/26:0196771 - isvavai.cz</a>

Výsledek na webu

<a href="https://aclanthology.org/2024.clinicalnlp-1.8/" target="_blank" >https://aclanthology.org/2024.clinicalnlp-1.8/</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.18653/v1/2024.clinicalnlp-1.8" target="_blank" >10.18653/v1/2024.clinicalnlp-1.8</a>

Jazyk výsledku

angličtina

Název v původním jazyce

DAIC-WOZ: On the Validity of Using the Therapist's prompts in Automatic Depression Detection from Clinical Interviews

Popis výsledku v původním jazyce

Automatic depression detection from conversational data has gained significant interest in recent years. The DAIC-WOZ dataset, interviews conducted by a human-controlled virtual agent, has been widely used for this task. Recent studies have reported enhanced performance when incorporating interviewer's prompts into the model. In this work, we hypothesize that this improvement might be mainly due to a bias present in these prompts, rather than the proposed architectures and methods. Through ablation experiments and qualitative analysis, we discover that models using interviewer's prompts learn to focus on a specific region of the interviews, where questions about past experiences with mental health issues are asked, and use them as discriminative shortcuts to detect depressed participants. In contrast, models using participant responses gather evidence from across the entire interview. Finally, to highlight the magnitude of this bias, we achieve a 0.90 F1 score by intentionally exploiting it, the highest result reported to date on this dataset using only textual information. Our findings underline the need for caution when incorporating interviewers' prompts into models, as they may inadvertently learn to exploit targeted prompts, rather than learning to characterize the language and behavior that are genuinely indicative of the patient's mental health condition.

Název v anglickém jazyce

DAIC-WOZ: On the Validity of Using the Therapist's prompts in Automatic Depression Detection from Clinical Interviews

Popis výsledku anglicky

Automatic depression detection from conversational data has gained significant interest in recent years. The DAIC-WOZ dataset, interviews conducted by a human-controlled virtual agent, has been widely used for this task. Recent studies have reported enhanced performance when incorporating interviewer's prompts into the model. In this work, we hypothesize that this improvement might be mainly due to a bias present in these prompts, rather than the proposed architectures and methods. Through ablation experiments and qualitative analysis, we discover that models using interviewer's prompts learn to focus on a specific region of the interviews, where questions about past experiences with mental health issues are asked, and use them as discriminative shortcuts to detect depressed participants. In contrast, models using participant responses gather evidence from across the entire interview. Finally, to highlight the magnitude of this bias, we achieve a 0.90 F1 score by intentionally exploiting it, the highest result reported to date on this dataset using only textual information. Our findings underline the need for caution when incorporating interviewers' prompts into models, as they may inadvertently learn to exploit targeted prompts, rather than learning to characterize the language and behavior that are genuinely indicative of the patient's mental health condition.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the 6th Clinical Natural Language Processing Workshop, ClinicalNLP@NAACL

ISBN

979-8-89176-109-4

ISSN

—

e-ISSN

—

Počet stran výsledku

9

Strana od-do

82-90

Název nakladatele

Association for Computational Linguistics

Místo vydání

Mexico City

Místo konání akce

Mexico City

Datum konání akce

21. 6. 2024

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001606850700008