Category-aware Collaborative Sequential Recommendation

Sequential recommendation is the task of predicting the next items for users based on their interaction history. Modeling the dependence of the next action on the past actions accurately is crucial to this problem. Moreover, sequential recommendation often faces serious sparsity of item-to-item transitions in a user's action sequence, which limits the practical utility of such solutions.