mlsanityData to 14 Oct 2026

Building Patient Journeys in Hebrew: A Language Model for Clinical Timeline Extraction

arXiv 2512.11502 ↗First posted 12 Dec 2025cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 20 of its 31 references that are in the map. First posted on arXiv 12 Dec 2025.

Builds on

20 of 31 references.

References · 20

  1. 2016Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translationcited by 1,475
  2. 2019SciBERT: A Pretrained Language Model for Scientific Textcited by 1,414
  3. 2020Med-BERT: pre-trained contextualized embeddings on large-scale structured electronic health records for disease predictioncited by 373
  4. 2018A Multi-Axis Annotation Scheme for Event Temporal Relationscited by 75
  5. 2024Unpacking Tokenization: Evaluating Text Compression and its Correlation with Model Performancecited by 61
  6. 2019Joint Event and Temporal Relation Extraction with Shared Representations and Structured Predictioncited by 53
  7. 2020ECONET: Effective Continual Pretraining of Language Models for Event Temporal Reasoningcited by 38
  8. 2023Parameter-Efficient Fine-Tuning of LLaMA for the Clinical Domaincited by 38
  9. 2021Adapt-and-Distill: Developing Small, Fast and Effective Pretrained Language Models for Domainscited by 36
  10. 2019An Improved Neural Baseline for Temporal Relation Extractioncited by 32
  11. 2023Efficiently Adapting Pretrained Language Models To New Languagescited by 28
  12. 2019Deep Structured Neural Network for Event Temporal Relation Extractioncited by 25
8 more
  1. 2022On the Domain Adaptation and Generalization of Pretrained Language Models: A Surveycited by 23
  2. 2020Clinical Temporal Relation Extraction with Probabilistic Soft Logic Regularization and Global Inferencecited by 20
  3. 2021AVocaDo: Strategy for Adapting Vocabulary to Downstream Domaincited by 20
  4. 2021Extracting Temporal Event Relation with Syntax-guided Graph Transformercited by 16
  5. 2022Extracting or Guessing? Improving Faithfulness of Event Temporal Relation Extractioncited by 15
  6. 2024Adapting LLMs to Hebrew: Unveiling DictaLM 2.0 with Enhanced Vocabulary and Instruction Capabilitiescited by 13
  7. 2024Will LLMs Replace the Encoder-Only Models in Temporal Relation Classification?cited by 6
  8. 2025Beyond Pairwise: Global Zero-shot Temporal Graph Generationcited by 3

Built on by

Not cited yet.