Deformable phrase level attention: A flexible approach for improving AI based medical coding
This work proposes a novel deformable, phrase-level attention mechanism that enhances text classification models in the extraction of medical concepts from clinical text documents. We demonstrate strong performances on two clinical text datasets and showcase real-world deployability of our method.
Authors
- John Gounley
- Drahomíra Herrmannová (ORCID: https://orcid.org/0000-0002-2730-1546)
- Shang Gao (ORCID: https://orcid.org/0000-0003-1803-1457)
- Heidi A. Hanson (ORCID: https://orcid.org/0000-0003-0056-196X)
- Christoph Metzner (ORCID: https://orcid.org/0000-0002-4051-620X)
Institutions
- Oak Ridge National Laboratory (US)
- RELX Group (Netherlands) (NL)
- RELX Group (United States) (US)
- Thomson Reuters (United States) (US)
- University of Tennessee at Knoxville (US)
Publication Details
- Journal
- Artificial Intelligence in Medicine
- Published
- 2025-11-20
- DOI
- https://doi.org/10.1016/j.artmed.2025.103299
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 0.00