Task-Feature Collaborative Learning with Application to Personalized Attribute Prediction

As an effective learning paradigm against insufficient training samples, Multi-task learning (MTL) encourages knowledge sharing across multiple related tasks so as to improve the overall performance. In MTL, a major challenge springs from the phenomenon that sharing the knowledge with dissimilar and hard tasks, known as negative transfer, often results in a worsened performance. Though a substantial amount of studies have been carried out against the negative transfer, most of the existing methods only model the transfer relationship at the task-level, with the transfer across features and tasks left unconsidered. Different from the existing methods, our goal is to alleviate negative transfer collaboratively at both the task- and feature-level. Specifically, we propose a novel multi-task learning method called Task-Feature Collaborative Learning (TFCL), leveraging the collaborative grouping of features and tasks while suppressing inter-group knowledge sharing. We then propose an optimization method for TFCL with a global convergence guarantee. Moreover, we show that the intermediate procedures of the method not only reveal how the transfer across features and tasks takes place, but also shed some light on the task-feature grouping effect. As a practical extension, we extend TFCL to the personalized attribute prediction problem with fine-grained modeling of user behaviors.