English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Electronic health records (EHR) contain narrative notes that provideextensive details on the medical condition and management of patients. Naturallanguage processing (NLP) of clinical notes can use observed frequencies ofclinical terms as predictive features for downstream applications such asclinical decision making and patient trajectory prediction. However, due to thevast number of highly similar and related clinical concepts, a more effectivemodeling strategy is to represent clinical terms as semantic embeddings viarepresentation learning and use the low dimensional embeddings as featurevectors for predictive modeling. To achieve efficient representation,fine-tuning pretrained language models with biomedical knowledge graphs maygenerate better embeddings for biomedical terms than those from standardlanguage models alone. These embeddings can effectively discriminate synonymouspairs of from those that are unrelated. However, they often fail to capturedifferent degrees of similarity or relatedness for concepts that arehierarchical in nature. To overcome this limitation, we propose HiPrBERT, anovel biomedical term representation model trained on additionally complieddata that contains hierarchical structures for various biomedical terms. Wemodify an existing contrastive loss function to extract information from thesehierarchies. Our numerical experiments demonstrate that HiPrBERT effectivelylearns the pair-wise distance from hierarchical information, resulting in asubstantially more informative embeddings for further biomedical applications

Hierarchical Pretraining for Biomedical Term Embeddings