Multi-interest learning method for sequential recommendation aims to predict the next item according to user multi-faceted interests given the user historical interactions. Existing methods mainly consist of two modules: the multi-interest extraction module that learns user multi-interest embeddings to capture the user multi-interests, and the multi-interest weight prediction module that learns the weight of each interest for aggregating the learned multi-interest embeddings to derive the user embedding, used for predicting the user rating to an item. Despite their effectiveness, existing methods have two key limitations: 1) they directly feed the user interactions into the two modules, while ignoring their different learning objectives, and 2) they merely consider the centrality of the user interactions to learn the user multi-interests, while overlooking their dispersion. To tackle these limitations, we propose a prompt-based multi-interest learning method (PoMRec), where specific prompts are inserted into user interactions to make them adaptive to different learning objectives of the two modules. Moreover, we utilize both the mean and variance embeddings of user interactions to derive the user multi-interest embeddings for comprehensively model the user multi-interests. We conduct extensive experiments on two public datasets, and the results verify that our proposed PoMRec outperforms the state-of-the-art multi-interest learning methods.

翻译：序列推荐中的多兴趣学习方法旨在根据用户历史交互行为预测其下一个交互项目，以捕捉用户的多方面兴趣。现有方法主要包含两个模块：多兴趣提取模块，用于学习用户的多兴趣嵌入以表征用户的多种兴趣；以及多兴趣权重预测模块，用于学习每个兴趣的权重，从而聚合已学习到的多兴趣嵌入生成用户嵌入，用于预测用户对项目的评分。尽管现有方法有效，但仍存在两个关键局限：1）直接将用户交互输入两个模块，忽略了两者不同的学习目标；2）仅考虑用户交互的集中性来学习用户的多兴趣，而未关注其离散性。为解决这些问题，我们提出一种基于提示的多兴趣学习方法（PoMRec），该方法在用户交互中插入特定提示，使其适应两个模块的不同学习目标。此外，我们同时利用用户交互的均值嵌入和方差嵌入来推导用户的多兴趣嵌入，以全面建模用户的多兴趣。在两个公开数据集上进行的广泛实验表明，所提出的PoMRec方法优于现有的最先进多兴趣学习方法。