This Monday (2021-07-12) marked the official start of the ADAM project, a collaboration between AZ Delta’s RADAR team and our own Internet and Data science Lab (IDLab, joint venture of UGent & imec).

Given I will be lead researcher on the Natural Language Understanding aspect of this project, I wanted to talk about it on my blog as well. Our shared goal in the ADAM project is to enable personalized medicine through deep patient understanding, which requires natural language processing of Dutch clinical notes.

In this technical report, the training procedure of our Transformer-based Dutch-to-English translation model will be outlined. This model was trained exclusively using freely available datasets, and yet achieves a performance level superior to the current state of the art in the domain.

We show this by comparing the precision and recall of MetaMap’s medical concepts extraction when executed on a set of Dutch clinical notes after translation to English, as well as by using more conventional quality metrics for translation models (BLEU…).