Custom NLP pipeline for information extraction in a clinical setting

State-of-the-art natural language processing tools are rapidly improving in being able to identify real-world objects in texts (e.g. person, location, organization). However, the vocabularies and document structures used in a medical setting can vary from general language corpus from which the NLP models were trained; resulting in poor performance.
This project will fine-tune existing NLP models to perform highly on medical text documents in order to recognise and extract pertinent information about a patient (e.g. diseases, medical findings) in a structured format. Domain-specific guidelines will be developed, using subject matter experts, to annotate training corpora to aid the training of clinical language NLP models.

Faculty Supervisor:

Panayiotis Pappas;Vered Shwartz

Student:

Partner:

iClinic Systems Inc.

Discipline:

Computer science

Sector:

Manufacturing; Professional, scientific and technical services

University:

Simon Fraser University; The University of British Columbia

Program: