Deep learning provides a great way to reach human baselines in classification problems given that enough annotated data is avaliable. My job at insight is to try to make it work by reducing drastically the expensive annotation work on data. I use and design label noise correction algorithms for image classification that use 10 times less annotations and obtain comparable results to fully annoated datasets.
I graduated a MSc degree from the French Engineering school CentraleSupelec in 2018 were I developed an extensive scientific background. The specialisation I chose was interactive systems, machine learning, data management and representation. I have been working as a research assistant in the Insight Center for Data Analytics @DCU since October 2018 under the supervision of Dr. Kevin McGuiness and Prof. Noel O'Connor.
I am interested in a broad array of applications for machine learning algorithms including computer vision and decision making.