Sequential domain-adaptive machine learning

Abstract

This poster recaps two collaboration projects I did during my time as Niels Stensen Fellow at the University of Copenhagen. The main point of my proposal was to study “sequential domain adaptation”. The left column describes an application to “spatial sequences”, i.e., multi-site biomedical imaging, and the right column describes an application to “temporal sequences”, i.e. natural language processing over data collected in snapshots over time. The main take-home message from the work in this Fellowship is that domain adaptation is a solution to training machine learning models under sampling bias and that sequential adaptation allows for updating.