arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.06388cs.IRcs.AI

HyMoERec:用于序列推荐的混合专家模型

HyMoERec: Hybrid Mixture-of-Experts for Sequential Recommendation

Kunrong Li, Zhu Sun, Kwan Hui Lim

AI总结:

HyMoERec 提出混合专家架构与自适应融合机制,解决序列推荐中用户行为和物品复杂性异质性问题,并在两个数据集上超越现有最优方法。

摘要:

我们提出 HyMoERec,一种新颖的序列推荐框架,旨在解决现有模型中统一的位置逐点前馈网络的局限性。当前方法平等对待所有用户交互和物品,忽视了用户行为模式的异质性以及物品复杂性的多样性。HyMoERec 首先引入一种混合专家架构,该架构结合共享专家分支和专用专家分支,并采用自适应专家融合机制用于序列推荐任务。该设计能够捕获不同用户和物品的多样化推理,同时确保训练稳定。在 MovieLens-1M 和 Beauty 数据集上的实验表明,HyMoERec 持续优于最先进的基线方法。

英文摘要:

We propose HyMoERec, a novel sequential recommendation framework that addresses the limitations of uniform Position-wise Feed-Forward Networks in existing models. Current approaches treat all user interactions and items equally, overlooking the heterogeneity in user behavior patterns and diversity in item complexity. HyMoERec initially introduces a hybrid mixture-of-experts architecture that combines shared and specialized expert branches with an adaptive expert fusion mechanism for the sequential recommendation task. This design captures diverse reasoning for varied users and items while ensuring stable training. Experiments on MovieLens-1M and Beauty datasets demonstrate that HyMoERec consistently outperforms state-of-the-art baselines.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑