CDRNP: Cross-Domain Recommendation to Cold-Start Users via Neural Process

被引:0
|
作者
Li, Xiaodong [1 ,2 ]
Sheng, Jiawei [1 ]
Cao, Jiangxia [3 ]
Zhang, Wenyuan [1 ,2 ]
Li, Quangang [1 ]
Liu, Tingwen [1 ,2 ]
机构
[1] Chinese Acad Sci, Inst Informat Engn, Beijing, Peoples R China
[2] UCAS, Sch Cyber Secur, Beijing, Peoples R China
[3] Kuaishou Inc, Beijing, Peoples R China
关键词
Cross-Domain Recommendation; Cold-Start Recommendation; Meta Learning; Neural Process;
D O I
10.1145/3616855.3635794
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Cross-domain recommendation (CDR) has been proven as a promising way to tackle the user cold-start problem, which aims to make recommendations for users in the target domain by transferring the user preference derived from the source domain. Traditional CDR studies follow the embedding and mapping (EMCDR) paradigm, which transfers user representations from the source to target domain by learning a user-shared mapping function, neglecting the user-specific preference. Recent CDR studies attempt to learn user-specific mapping functions in meta-learning paradigm, which regards each user's CDR as an individual task, but neglects the preference correlations among users, limiting the beneficial information for user representations. Moreover, both of the paradigms neglect the explicit user-item interactions from both domains during the mapping process. To address the above issues, this paper proposes a novel CDR framework with neural process (NP), termed as CDRNP. Particularly, it develops the meta-learning paradigm to leverage user-specific preference, and further introduces a stochastic process by NP to capture the preference correlations among the overlapping and cold-start users, thus generating more powerful mapping functions by mapping the user-specific preference and common preference correlations to a predictive probability distribution. In addition, we also introduce a preference remainer to enhance the common preference from the overlapping users, and finally devises an adaptive conditional decoder with preference modulation to make prediction for cold-start users with items in the target domain. Experimental results demonstrate that CDRNP outperforms previous SOTA methods in three real-world CDR scenarios.
引用
收藏
页码:378 / 386
页数:9
相关论文
共 50 条
  • [1] Cross-Domain Recommendation To Cold-Start Users Via Categorized Preference Transfer
    Liu, Xiaoyang
    Fu, Xiaoyang
    De Meo, Pasquale
    Fiumara, Giacomo
    COMPUTER JOURNAL, 2024,
  • [2] Cross-Domain Recommendation To Cold-Start Users Via Categorized Preference Transfer
    Liu, Xiaoyang
    Fu, Xiaoyang
    De Meo, Pasquale
    Fiumara, Giacomo
    Computer Journal, 1600, 67 (08): : 2610 - 2621
  • [3] Cross-Domain Recommendation to Cold-Start Users via Variational Information Bottleneck
    Cao, Jiangxia
    Sheng, Jiawei
    Cong, Xin
    Liu, Tingwen
    Wang, Bin
    2022 IEEE 38TH INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE 2022), 2022, : 2209 - 2223
  • [4] Cross-Domain Recommendation for Cold-Start Users via Neighborhood Based Feature Mapping
    Wang, Xinghua
    Peng, Zhaohui
    Wang, Senzhang
    Yu, Philip S.
    Fu, Wenjing
    Hong, Xiaoguang
    DATABASE SYSTEMS FOR ADVANCED APPLICATIONS, DASFAA 2018, PT I, 2018, 10827 : 158 - 165
  • [5] CDLFM: cross-domain recommendation for cold-start users via latent feature mapping
    Wang, Xinghua
    Peng, Zhaohui
    Wang, Senzhang
    Yu, Philip S.
    Fu, Wenjing
    Xu, Xiaokang
    Hong, Xiaoguang
    KNOWLEDGE AND INFORMATION SYSTEMS, 2020, 62 (05) : 1723 - 1750
  • [6] CATN: Cross-Domain Recommendation for Cold-Start Users via Aspect Transfer Network
    Zhao, Cheng
    Li, Chenliang
    Xiao, Rong
    Deng, Hongbo
    Sun, Aixin
    PROCEEDINGS OF THE 43RD INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL (SIGIR '20), 2020, : 229 - 238
  • [7] CDLFM: cross-domain recommendation for cold-start users via latent feature mapping
    Xinghua Wang
    Zhaohui Peng
    Senzhang Wang
    Philip S. Yu
    Wenjing Fu
    Xiaokang Xu
    Xiaoguang Hong
    Knowledge and Information Systems, 2020, 62 : 1723 - 1750
  • [8] Improving Top-N Recommendation for Cold-Start Users via Cross-Domain Information
    Mirbakhsh, Nima
    Ling, Charles X.
    ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA, 2015, 9 (04) : 1 - 19
  • [9] Transfer-Meta Framework for Cross-domain Recommendation to Cold-Start Users
    Zhu, Yongchun
    Ge, Kaikai
    Zhuang, Fuzhen
    Xie, Ruobing
    Xi, Dongbo
    Zhang, Xu
    Lin, Leyu
    He, Qing
    SIGIR '21 - PROCEEDINGS OF THE 44TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL, 2021, : 1813 - 1817
  • [10] Semi-Supervised Learning for Cross-Domain Recommendation to Cold-Start Users
    Kang, SeongKu
    Hwang, Junyoung
    Lee, Dongha
    Yu, Hwanjo
    PROCEEDINGS OF THE 28TH ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT (CIKM '19), 2019, : 1563 - 1572