Biotech & healthResearch findings
A Transfer Learning Approach for Predicting Entity Modifiers in Clinical Text for Opioid Use Disorder Detection
Original published 2024.01.27 07:19 KSTAdded to Blackdeck 2026.11.01 18:30 KST
What has changed?
A multi-task transformer architecture jointly learns and predicts entity modifiers from shared and novel clinical contexts
Key points
- The semantics of clinical entities are significantly altered by modifiers such as negation and severity
- Existing models train independently for each modifier using regular expressions or feature weights
- A multi-task transformer jointly learns modifiers from shared and OUD-specific clinical data
Connect it to your work
The model handles both shared modifiers from SemEval and novel opioid-specific modifiers
What remains unproven
This research note is based on an abstract; check the original for test conditions and application scope.
Originals used for this note
Links open the original publisher in a new tab.