Sequential recommendation aims to recommend the next item that matches a user's

Most existing work in this area relies on non-probabilistic approaches or on complex Reinforcement Learning (RL)-based hard structures.