Computer Engineering and Applications ›› 2023, Vol. 59 ›› Issue (10): 104-113.DOI: 10.3778/j.issn.1002-8331.2202-0128

• Pattern Recognition and Artificial Intelligence • Previous Articles     Next Articles

Compound Convolutional and Self-Attention Network for Session-Based Recommendation

XIAO Yan, HUO Lin   

  1. College of Computer and Electronic Information, Guangxi University, Nanning 530000, China
  • Online:2023-05-15 Published:2023-05-15

基于复合卷积和自注意力的会话推荐

肖妍,霍林   

  1. 广西大学 计算机与电子信息学院,南宁 530000

Abstract: Recently, methods based on convolutional neural networks(CNN) have shown potential in modeling conversational data, especially in extracting complex local behavioral interactions of users. Recurrent neural networks(RNNs) have difficulty for learning item dependencies from a distance, while self-attention(SA) structures have the ability to model the sequence of conversational events and capture interactions between distant items, which is a superior choice for session data. Therefore, a compound convolutional and self-attention network(CCNN-SA) architecture is proposed, which utilizes the complex local features extracted by two convolutional modules, and uses a multi-head self-attention structure to learn long-term interactions from conversational events, and this flexible and unified network architecture facilitates comprehensive modeling of various important features of conversational sequences. The proposed model is validated on two benchmark datasets from e-commerce, and the evaluation metrics Recall@20 and MRR@20, which characterize hit rate and prediction result ranking, are improved by 2.79% and 5.87% respectively on the YOOCHOOSE dataset, are improved by 2.17% and 6.43% respectively on the DIGINETIC dataset, which verifies the validity and rationality of the model.

Key words: compound convolution, multi-head self-attention, session-based recommendation

摘要: 最近,基于卷积神经网络(convolutional neural networks,CNN)的方法显示了在会话数据建模中的潜力,尤其是在提取用户复杂的局部行为交互方面。递归神经网络(recurrent neural networks,RNN)很难从远距离学习项目依赖,而自注意力(self-attention,SA)结构具有对会话事件的顺序进行建模并捕获远距离项目之间交互关系的能力,对于会话数据来说是一种优越的选择。因此提出了一种复合卷积自注意力架构(compound convolutional and self-attention network,CCNN-SA)。该架构利用了两个卷积模块提取的复杂局部特征,同时使用了多头自注意力结构从会话事件中学习长期交互关系,这种灵活而统一的网络体系结构有利于对会话序列的各种重要特征进行综合建模。在两个来自电子商务的基准数据集上对所提出的模型进行了验证,在表征命中率和预测结果排名的评估指标Recall@20和MRR@20上,对于YOOCHOOSE数据集分别提升了2.79%和5.87%,在DIGINETIC数据集上分别提升了2.17%和6.43%,验证了模型的有效性与合理性。

关键词: 复合卷积, 多头自注意力, 会话推荐