Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2512.11502 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915671170875392 |
|---|---|
| author | Hashiloni, Kai Golan Nokai, Brenda Kasabe Shevach, Michal Shemesh, Esthy Bartin, Ronit Bergrin, Anna Harel, Liran Dershowitz, Nachum Arad, Liat Nadai Bar, Kfir |
| author_facet | Hashiloni, Kai Golan Nokai, Brenda Kasabe Shevach, Michal Shemesh, Esthy Bartin, Ronit Bergrin, Anna Harel, Liran Dershowitz, Nachum Arad, Liat Nadai Bar, Kfir |
| contents | We present a new Hebrew medical language model designed to extract structured clinical timelines from electronic health records, enabling the construction of patient journeys. Our model is based on DictaBERT 2.0 and continually pre-trained on over five million de-identified hospital records. To evaluate its effectiveness, we introduce two new datasets -- one from internal medicine and emergency departments, and another from oncology -- annotated for event temporal relations. Our results show that our model achieves strong performance on both datasets. We also find that vocabulary adaptation improves token efficiency and that de-identification does not compromise downstream performance, supporting privacy-conscious model development. The model is made available for research use under ethical restrictions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_11502 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Building Patient Journeys in Hebrew: A Language Model for Clinical Timeline Extraction Hashiloni, Kai Golan Nokai, Brenda Kasabe Shevach, Michal Shemesh, Esthy Bartin, Ronit Bergrin, Anna Harel, Liran Dershowitz, Nachum Arad, Liat Nadai Bar, Kfir Computation and Language We present a new Hebrew medical language model designed to extract structured clinical timelines from electronic health records, enabling the construction of patient journeys. Our model is based on DictaBERT 2.0 and continually pre-trained on over five million de-identified hospital records. To evaluate its effectiveness, we introduce two new datasets -- one from internal medicine and emergency departments, and another from oncology -- annotated for event temporal relations. Our results show that our model achieves strong performance on both datasets. We also find that vocabulary adaptation improves token efficiency and that de-identification does not compromise downstream performance, supporting privacy-conscious model development. The model is made available for research use under ethical restrictions. |
| title | Building Patient Journeys in Hebrew: A Language Model for Clinical Timeline Extraction |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2512.11502 |