1 Introduction
Human can learn multiple tasks simultaneously and during this learning process, human can use the knowledge learned in a task to help the learning of another task. For example, according to our experience in learning to play tennis and squash, we find that the skill of playing tennis can help learn to play squash and vice versa. Inspired by such human learning ability, Multi-Task Learning (MTL) [1], a learning paradigm in machine learning, aims to learn multiple related tasks jointly so that the knowledge contained in a task can be leveraged by other tasks, with the hope of improving the generalization performance of all the tasks at hand.