A multi-task learning example for the paper https://arxiv.org/abs/1705.07115 - View it on GitHub