DS@GT at eRisk 2025: From prompts to predictions, benchmarking early depression detection with conversational agent based assessments and temporal attention models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915390058135552 |
|---|---|
| author | Miyaguchi, Anthony Guecha, David Chiu, Yuwen Gaur, Sidharth |
| author_facet | Miyaguchi, Anthony Guecha, David Chiu, Yuwen Gaur, Sidharth |
| contents | This Working Note summarizes the participation of the DS@GT team in two eRisk 2025 challenges. For the Pilot Task on conversational depression detection with large language-models (LLMs), we adopted a prompt-engineering strategy in which diverse LLMs conducted BDI-II-based assessments and produced structured JSON outputs. Because ground-truth labels were unavailable, we evaluated cross-model agreement and internal consistency. Our prompt design methodology aligned model outputs with BDI-II criteria and enabled the analysis of conversational cues that influenced the prediction of symptoms. Our best submission, second on the official leaderboard, achieved DCHR = 0.50, ADODL = 0.89, and ASHR = 0.27. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_10958 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DS@GT at eRisk 2025: From prompts to predictions, benchmarking early depression detection with conversational agent based assessments and temporal attention models Miyaguchi, Anthony Guecha, David Chiu, Yuwen Gaur, Sidharth Computation and Language This Working Note summarizes the participation of the DS@GT team in two eRisk 2025 challenges. For the Pilot Task on conversational depression detection with large language-models (LLMs), we adopted a prompt-engineering strategy in which diverse LLMs conducted BDI-II-based assessments and produced structured JSON outputs. Because ground-truth labels were unavailable, we evaluated cross-model agreement and internal consistency. Our prompt design methodology aligned model outputs with BDI-II criteria and enabled the analysis of conversational cues that influenced the prediction of symptoms. Our best submission, second on the official leaderboard, achieved DCHR = 0.50, ADODL = 0.89, and ASHR = 0.27. |
| title | DS@GT at eRisk 2025: From prompts to predictions, benchmarking early depression detection with conversational agent based assessments and temporal attention models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2507.10958 |