@inproceedings{roll-etal-2026-team,
    title = "Team Aurevia at {CLP}sych 2026: Local Healthcare {NLP} for Schema-Constrained Self-State Modeling",
    author = "Roll, Nathan  and
      Yi, Irene  and
      Aldogom, Sufian  and
      Brown, Grace  and
      Basile, Eric  and
      Gutterman, Isaac  and
      Tennakoon, Lakshika  and
      Ahmed, Ammar",
    editor = "Zirikly, Aya  and
      Bar, Kfir  and
      MacAvaney, Sean  and
      Ireland, Molly  and
      Ophir, Yaakov  and
      Atzil-Slonim, Dana  and
      Varadarajan, Vasudha  and
      Bedrick, Steven  and
      Desmet, Bart",
    booktitle = "Proceedings of the 10th Workshop on Computational Linguistics and Clinical Psychology ({CLP}sych 2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.clpsych-1.43/",
    pages = "531--534",
    ISBN = "979-8-89176-421-7",
    abstract = "Team Aurevia introduces a local open-weight healthcare NLP system for the CLPsych 2026 Shared Task, predicting MIND-coded self-state elements, moments of change, summaries, anddynamic signatures from social media timelines. The task is difficult because coarse presence, fine-grained ABCD subelements, and timeline-level change require different longitudinal evidence over privacy-sensitive mental-health language. Our system combines TF-IDF retrieval, schema-constrained local Qwen2.5 prompting, ordinal calibration, and conservative post-processing. Among official runs, Aurevia ranked 3rd of 17 for Task 1.2 presence prediction, 5th of 13 overall for Task 3.1, 1st on Task 3.1 consistency, and 2nd of 9 for MIND-coded deterioration signatures, showing that constrained local LLM pipelines can remain competitive in sensitive healthcare NLP while reducing reliance on hosted proprietary inference."
}