Code and data used to conduct an empirical analysis of popular regularised multi-task learning methods.
This project includes the experimental setting for testing different multi-task learning methods implemented in MALSAR 1.1, as well as an implementation of "Regularized Multi–Task Learning, T. Evgeniou and M. Pontil (2004)" with gradient descent optimisation.
MALSAR 1.1 https://github.com/jiayuzhou/MALSAR