CareLab at #SMM4H-HeaRD 2025: Insomnia Detection and Food Safety Event Extraction with Domain-Aware Transformers

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liang, Zihan, Pan, Ziwen, Dey, Sumon Kanti, Ismail, Azra
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909656577736704
author Liang, Zihan
Pan, Ziwen
Dey, Sumon Kanti
Ismail, Azra
author_facet Liang, Zihan
Pan, Ziwen
Dey, Sumon Kanti
Ismail, Azra
contents This paper presents our system for the SMM4H-HeaRD 2025 shared tasks, specifically Task 4 (Subtasks 1, 2a, and 2b) and Task 5 (Subtasks 1 and 2). Task 4 focused on detecting mentions of insomnia in clinical notes, while Task 5 addressed the extraction of food safety events from news articles. We participated in all subtasks and report key findings across them, with particular emphasis on Task 5 Subtask 1, where our system achieved strong performance-securing first place with an F1 score of 0.958 on the test set. To attain this result, we employed encoder-based models (e.g., RoBERTa), alongside GPT-4 for data augmentation. This paper outlines our approach, including preprocessing, model architecture, and subtask-specific adaptations
format Preprint
id arxiv_https___arxiv_org_abs_2506_18185
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CareLab at #SMM4H-HeaRD 2025: Insomnia Detection and Food Safety Event Extraction with Domain-Aware Transformers
Liang, Zihan
Pan, Ziwen
Dey, Sumon Kanti
Ismail, Azra
Computation and Language
Artificial Intelligence
This paper presents our system for the SMM4H-HeaRD 2025 shared tasks, specifically Task 4 (Subtasks 1, 2a, and 2b) and Task 5 (Subtasks 1 and 2). Task 4 focused on detecting mentions of insomnia in clinical notes, while Task 5 addressed the extraction of food safety events from news articles. We participated in all subtasks and report key findings across them, with particular emphasis on Task 5 Subtask 1, where our system achieved strong performance-securing first place with an F1 score of 0.958 on the test set. To attain this result, we employed encoder-based models (e.g., RoBERTa), alongside GPT-4 for data augmentation. This paper outlines our approach, including preprocessing, model architecture, and subtask-specific adaptations
title CareLab at #SMM4H-HeaRD 2025: Insomnia Detection and Food Safety Event Extraction with Domain-Aware Transformers
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2506.18185