تقرير
A Comparative Study of Pretrained Language Models for Long Clinical Text
العنوان: | A Comparative Study of Pretrained Language Models for Long Clinical Text |
---|---|
المؤلفون: | Li, Yikuan, Wehbe, Ramsey M., Ahmad, Faraz S., Wang, Hanyin, Luo, Yuan |
سنة النشر: | 2023 |
المجموعة: | Computer Science |
مصطلحات موضوعية: | Computer Science - Computation and Language, Computer Science - Artificial Intelligence |
الوصف: | Objective: Clinical knowledge enriched transformer models (e.g., ClinicalBERT) have state-of-the-art results on clinical NLP (natural language processing) tasks. One of the core limitations of these transformer models is the substantial memory consumption due to their full self-attention mechanism, which leads to the performance degradation in long clinical texts. To overcome this, we propose to leverage long-sequence transformer models (e.g., Longformer and BigBird), which extend the maximum input sequence length from 512 to 4096, to enhance the ability to model long-term dependencies in long clinical texts. Materials and Methods: Inspired by the success of long sequence transformer models and the fact that clinical notes are mostly long, we introduce two domain enriched language models, Clinical-Longformer and Clinical-BigBird, which are pre-trained on a large-scale clinical corpus. We evaluate both language models using 10 baseline tasks including named entity recognition, question answering, natural language inference, and document classification tasks. Results: The results demonstrate that Clinical-Longformer and Clinical-BigBird consistently and significantly outperform ClinicalBERT and other short-sequence transformers in all 10 downstream tasks and achieve new state-of-the-art results. Discussion: Our pre-trained language models provide the bedrock for clinical NLP using long texts. We have made our source code available at https://github.com/luoyuanlab/Clinical-Longformer, and the pre-trained models available for public download at: https://huggingface.co/yikuan8/Clinical-Longformer. Conclusion: This study demonstrates that clinical knowledge enriched long-sequence transformers are able to learn long-term dependencies in long clinical text. Our methods can also inspire the development of other domain-enriched long-sequence transformers. Comment: arXiv admin note: substantial text overlap with arXiv:2201.11838 |
نوع الوثيقة: | Working Paper |
DOI: | 10.1093/jamia/ocac225 |
URL الوصول: | http://arxiv.org/abs/2301.11847 |
رقم الأكسشن: | edsarx.2301.11847 |
قاعدة البيانات: | arXiv |
DOI: | 10.1093/jamia/ocac225 |
---|