6 researchers propose a controllable multi-interest framework for recommendation, which is used to recommend items based on a user's click sequence.
