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

Objective: To comparatively evaluate several transformer model architecturesat identifying protected health information (PHI) in the i2b2/UTHealth 2014clinical text de-identification challenge corpus.  Methods: The i2b2/UTHealth 2014 corpus contains N=1304 clinical notesobtained from N=296 patients. Using a transfer learning framework, we fine-tuneseveral transformer model architectures on the corpus, including: BERT-base,BERT-large, ROBERTA-base, ROBERTA-large, ALBERT-base and ALBERT-xxlarge. Duringfine-tuning we vary the following model hyper-parameters: batch size, numbertraining epochs, learning rate and weight decay. We fine tune models on atraining data set, we evaluate and select optimally performing models on anindependent validation dataset, and lastly assess generalization performance ona held-out test dataset. We assess model performance in terms of accuracy,precision (positive predictive value), recall (sensitivity) and F1 score(harmonic mean of precision and recall). We are interested in overall modelperformance (PHI identified vs. PHI not identified), as well as PHI-specificmodel performance.  Results: We observe that the ROBERTA-large models perform best at identifyingPHI in the i2b2/UTHealth 2014 corpus, achieving >99% overall accuracy and 96.7%recall/precision on the heldout test corpus. Performance was good across manyPHI classes; however, accuracy/precision/recall decreased for identification ofthe following entity classes: professions, organizations, ages, and certainlocations.  Conclusions: Transformers are a promising model class/architecture forclinical text de-identification. With minimal hyper-parameter tuningtransformers afford researchers/clinicians the opportunity to obtain (near)state-of-the-art performance.

A Comparative Evaluation Of Transformer Models For De-Identification Of Clinical Text Data