Understanding patient complaint characteristics using contextual clinical BERT embeddings (42nd IEEE EMBC 2020)
To understand patient complaints in conversational applications, we propose a two-fold approach using Word2vec and BERT to detect the characterizations of entities like symptoms presented by general users in contexts where they would describe their symptoms to a clinician.
Read MoreContinuous Improvement of Medical Diagnostic Systems with Large Scale Patient Vignette Simulation (29th ACM CIKM 2020)
To address continuous large-scale evaluation of diagnostic systems, we proposed a novel patient vignette simulation algorithm within an iterative clinician-in-the-loop methodology for semi-automatically evaluating and deploying medical diagnostic systems in production.
Read MoreA Dictionary-based Oversampling Approach to Clinical Document Classification on Small and Imbalanced Dataset (19th IEEE/ACM WI-IAT 2020)
To address data imbalance in medical document classification, we propose a probabilistic dictionary-based data augmentation approach by oversampling on the minority class and creating new documents with high variety by using synonyms’ similarities with the original medical entity term.
Read MorePredicting missing and noisy links via neighbourhood preserving graph embeddings in a clinical knowlegebase (19th IEEE ICMLA 2020)
For link discovery and quality assessments of medical knowledge graphs, we propose a novel neighborhood – based relationship entity embedding algorithm to simultaneously predict both noisy and missing links between medical entities such as diseases and symptoms.
Read More