Key points are not available for this paper at this time.
It is often desirable to model multiple objectives in real-world web applications, such as user satisfaction and user engagement in recommender systems. Multi-task learning has become the standard approach for such applications recently.
Qin et al. (Thu,) studied this question.