This project encompasses the development of a novel end-to-end automated framework for literature extraction of machine learning applied in minimally invasive surgery by training the state-of-the-art Bidirectional Encoder Representations from Transformers (BERT) language model. A knowledge graph generated nodes and relationships through word embeddings and graph algorithms. Annotations were be made semi-autonomously and the system was designed as a web application to be readily deployed in a clinical setting. The proposed framework can be used as a diagnosis and procedural assistant or used to accelerate literature and systematic reviews.

Infos

Participants

• Vanier College

Année : 2021

Région : MONTREAL REGIONAL SCIENCE & TECHNOLOGY FAIR

Type de projet : E

Niveau scolaire : Collégial 1

École : Vanier College

Prix et distinctions

Finale Québécoise

MONTREAL REGIONAL SCIENCE & TECHNOLOGY FAIR

• Prix de l’Association Mathématique du Québec

Photo