MLlm-DR: Towards Explainable Depression Recognition with MultiModal Large Language Models DOI: XXXXXXX.XXXXXXX Journal: JACM Volume: 37 4 131 8 CCS: Computing methodologies Artificial intelligence CCS: Human-centered computing Human-centered computing

Wei Zhang, Juan Chen, En Zhu, Wenhong Cheng, YunPeng Li, Yanbo J. Wang · 2025

Read MLlm-DR: Towards Explainable Depression Recognition with MultiModal Large Language Models DOI: XXXXXXX.XXXXXXX Journal: JACM Volume: 37 4 131 8 CCS: Computing methodologies Artificial intelligence CCS: Human-centered computing Human-centered computing by Wei Zhang, Juan Chen, En Zhu, Wenhong Cheng, YunPeng Li, Yanbo J. Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-05591.txt.

Description

Wei Zhang and co-authors build a multimodal language model that diagnoses depression from interviews and explains itself. Their 2025 paper addresses a trust gap: automated systems that score depression from video usually give no reason for the score, which limits clinical use. Their design pairs a smaller language model with a light query module that draws the needed evidence from the video and sound.

The system is trained to read interview material directly, since general multimodal models perform poorly on such data out of the box. An explained score lets a clinician check the reasoning rather than accept a number. The authors position the tool as decision support, not a replacement for judgment.

Catalog record

CREATORWei Zhang, Juan Chen, En Zhu, Wenhong Cheng, YunPeng Li, Yanbo J. Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-05591

Browse the full library · About the Tower of Babel