Journal of Frontiers of Computer Science and Technology

• Science Researches •     Next Articles

A Review of Embedding Methods and Domain Alignment Techniques in Dual Target Cross Domain Recommendation

HU Siyu,  MEI Hongyan,  YANG Haiyan,  CHENG Nai,  ZHANG Xiaoyu   

  1. School of Electronic and Information Engineering, Liaoning University of Technology, Jinzhou, Liaoning 121001, China

双目标跨域推荐中嵌入方法与领域对齐技术研究综述

胡思雨,梅红岩,杨海燕,程耐,张晓宇   

  1. 辽宁工业大学 电子与信息工程学院,辽宁 锦州 121001

Abstract: As a key branch of cross domain recommendation technology, dual target cross domain recommendation improves the efficiency of source domain and target domain recommendation synchronously by virtue of a two-way collaborative optimization mechanism, and is widely used in e-commerce, video distribution, news and information and other fields. This paper first introduces the domain hierarchy and overlapping scene characteristics in the cross-domain recommendation, and from the perspective of knowledge embedding methods, elaborates the core principles of collaborative filtering embedding, graph embedding, and self-monitoring learning embedding, and compares and analyzes their technical characteristics and applicable scenarios; From the perspective of domain alignment technology, four mainstream domain alignment schemes based on feature mapping, decoupled representation learning, meta-learning and federated learning are emphatically compared, and their technical differences and practical values are summarized. Secondly, the main data sets and evaluation indicators in the dual objective cross-domain recommendation are systemically sorted out, and the adaptation criteria of each data set and indicator are defined in combination with the characteristics of different cross-domain scenarios. Finally, based on the current research status and technical challenges, the future development direction of dual objective cross-domain recommendation is prospectively prospected.

Key words: dual target cross-domain recommendation, embedding method, domain alignment

摘要: 双目标跨域推荐作为跨域推荐技术的关键分支,凭借双向协同优化机制同步提升源域与目标域推荐效能,在电子商务、视频分发、新闻资讯等领域应用广泛。本文首先介绍了跨域推荐中的领域层次结构与重叠场景特征,并从知识嵌入方式方法角度,详细阐述协同过滤嵌入、图嵌入和自监督学习嵌入的核心原理,对比分析了其技术特性与适用场景;从领域对齐技术角度,着重对比了基于特征映射、解耦表示学习、元学习及联邦学习的四类主流领域对齐方案,总结了其技术差异与实践价值。其次,系统梳理了双目标跨域推荐中的主流数据集与评估指标,结合不同跨域场景特性,明确各数据集与指标的适配准则。最后,基于当前研究现状与技术挑战,对双目标跨域推荐的未来发展方向进行前瞻性展望。

关键词: 双目标跨域推荐, 嵌入方法, 领域对齐