The growing interest in biomedical and clinical research has led to a wide need of analyzing and understanding text in these domains. While today’s open-source NLP tools have integrated sophisticated neural architectures that improve their performance on general-domain text, they often lack convenient support for the analysis of biomedical text at the same level of accuracy.
In this talk, I will talk about the out-of-the-box biomedical and clinical packages in the Stanza Python NLP toolkit.
I will start by talking about the fully neural architectural design of Stanza, which allows it to generalize with ease to over 70 languages and multiple domains.
Then I will talk about how we extend this design to build and evaluate the biomedical and clinical pipelines for Stanza, which provide near state-of-the-art performance for linguistic analysis and entity recognition tasks.
Lastly, I will showcase how these biomedical and clinical models can be used in common research and text analysis scenarios. You can try out an online demo of these packages at: http://stanza.run/bio.