Camouflaged Variational Graph AutoEncoder Against Attribute Inference Attacks for Cross-Domain Recommendation

Cross-domain recommendation (CDR) aims to alleviate the data sparsity problem by leveraging the benefits of modeling two domains. However, existing research often focuses on the recommendation performance while ignores the privacy leakage issue. We find that an attacker can infer user attribute info...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE transactions on knowledge and data engineering Vol. 37; no. 7; pp. 3916 - 3932
Main Authors:	Xiong, Yudi, Guo, Yongxin, Pan, Weike, Yang, Qiang, Ming, Zhong, Zhang, Xiaojin, Han, Yu, Lin, Tao, Tang, Xiaoying
Format:	Journal Article
Language:	English
Published:	IEEE 01.07.2025
Subjects:	Accuracy attribute inference attacks Autoencoders Computational modeling cross-domain recommendation (CDR) Data models Graph convolutional networks Information leakage Knowledge transfer Privacy Protection recommendation system Recommender systems Variational autoencoders
ISSN:	1041-4347, 1558-2191
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	Cross-domain recommendation (CDR) aims to alleviate the data sparsity problem by leveraging the benefits of modeling two domains. However, existing research often focuses on the recommendation performance while ignores the privacy leakage issue. We find that an attacker can infer user attribute information from the knowledge (e.g., user preferences) transferred between the source and target domains. For example, in our experiments, the average inference accuracies of attack models on gender and age attributes are 0.8323 and 0.3897. The best-performing attack model achieves accuracies of 0.8847 and 0.4634, exceeding a random inference by 25.10% and 64.04%. We can see that the leakage of user attribute information may significantly exceed what would be expected from random inference. In this paper, we propose a novel recommendation framework named CVGAE (short for camouflaged variational graph autoencoder), which effectively models user behaviors and mitigates the risk of user attribute information leakage at the same time. Specifically, our CVGAE combines the strengths of VAEs in capturing latent features and variability with the ability of GCNs in exploiting high-order relational information. Moreover, to ensure against attribute inference attacks without sacrificing the recommendation performance, we design a user attribute protection module that fuses user attribute-camouflaged information with knowledge transfer during cross-domain processes. We then conduct extensive experiments on three real-world datasets, and find our CVGAE is able to achieve strong privacy protection while making little sacrifices in recommendation accuracy.
AbstractList	Cross-domain recommendation (CDR) aims to alleviate the data sparsity problem by leveraging the benefits of modeling two domains. However, existing research often focuses on the recommendation performance while ignores the privacy leakage issue. We find that an attacker can infer user attribute information from the knowledge (e.g., user preferences) transferred between the source and target domains. For example, in our experiments, the average inference accuracies of attack models on gender and age attributes are 0.8323 and 0.3897. The best-performing attack model achieves accuracies of 0.8847 and 0.4634, exceeding a random inference by 25.10% and 64.04%. We can see that the leakage of user attribute information may significantly exceed what would be expected from random inference. In this paper, we propose a novel recommendation framework named CVGAE (short for camouflaged variational graph autoencoder), which effectively models user behaviors and mitigates the risk of user attribute information leakage at the same time. Specifically, our CVGAE combines the strengths of VAEs in capturing latent features and variability with the ability of GCNs in exploiting high-order relational information. Moreover, to ensure against attribute inference attacks without sacrificing the recommendation performance, we design a user attribute protection module that fuses user attribute-camouflaged information with knowledge transfer during cross-domain processes. We then conduct extensive experiments on three real-world datasets, and find our CVGAE is able to achieve strong privacy protection while making little sacrifices in recommendation accuracy.
Author	Tang, Xiaoying Xiong, Yudi Lin, Tao Guo, Yongxin Yang, Qiang Zhang, Xiaojin Han, Yu Pan, Weike Ming, Zhong
Author_xml	– sequence: 1 givenname: Yudi orcidid: 0009-0001-3005-8225 surname: Xiong fullname: Xiong, Yudi email: xiongyudi2023@email.szu.edu.cn organization: College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, China – sequence: 2 givenname: Yongxin surname: Guo fullname: Guo, Yongxin email: 220019093@link.cuhk.edu.cn organization: School of Science and Engineering, Chinese University of Hong Kong (Shenzhen), Shenzhen, China – sequence: 3 givenname: Weike orcidid: 0000-0001-6326-9531 surname: Pan fullname: Pan, Weike email: panweike@szu.edu.cn organization: College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, China – sequence: 4 givenname: Qiang orcidid: 0000-0001-5059-8360 surname: Yang fullname: Yang, Qiang email: qyang@cse.ust.hk organization: Department of Computer Science and Engineering, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China – sequence: 5 givenname: Zhong orcidid: 0000-0001-9310-3460 surname: Ming fullname: Ming, Zhong email: mingz@szu.edu.cn organization: College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, China – sequence: 6 givenname: Xiaojin orcidid: 0000-0001-9065-6852 surname: Zhang fullname: Zhang, Xiaojin email: xiaojinzhang@hust.edu.cn organization: School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan, China – sequence: 7 givenname: Yu surname: Han fullname: Han, Yu email: han.yu@ntu.edu.sg organization: School of Computer Science and Engineering, Nanyang Technological University, Singapore – sequence: 8 givenname: Tao surname: Lin fullname: Lin, Tao email: lintao@westlake.edu.cn organization: School of Engineering, Westlake University, Hangzhou, China – sequence: 9 givenname: Xiaoying orcidid: 0000-0003-3955-1195 surname: Tang fullname: Tang, Xiaoying email: tangxiaoying@cuhk.edu.cn organization: School of Science and Engineering, Chinese University of Hong Kong (Shenzhen), Shenzhen, China
BookMark	eNpNkM1OwzAQhC1UJNrCAyBx8Auk-De2j1FaSkUlJFS4Rq6zKYHGrpzkwNuT0B647K5WM6PRN0MTHzwgdE_JglJiHncvy9WCESYXXKZSGX6FplRKnTBq6GS4iaCJ4ELdoFnbfhFCtNJ0iprcNqGvjvYAJf6wsbZdHbw94nW0p0-c9V1YeRdKiDg72Nq3Hc66Ltb7vgO88RVE8A7Gn3XfLa5CxHkMbZssQzPI8Ru40DTgy7_cW3Rd2WMLd5c9R-9Pq13-nGxf15s82yaOMdIlWihZpqCoEmWaKieZckoqLY2RLOUKhBCcEl3uK6pLrYHtHUkZDNMQIyifI3rOdWOXCFVxinVj409BSTHyKkZexciruPAaPA9nTw0A__RGE54K_gvEO2k_
CODEN	ITKEEH
Cites_doi	10.1145/3038912.3052569 10.1109/ICDM58522.2023.00053 10.1109/TKDE.2023.3295601 10.1007/s13278-020-0626-2 10.1145/2724720 10.48550/arXiv.1205.2618 10.1073/pnas.1218772110 10.1109/TKDE.2023.3253168 10.1016/j.dss.2015.03.008 10.1145/1526709.1526899 10.1145/3616855.3635830 10.1145/3637528.3671789 10.1145/3038912.3052695 10.24963/ijcai.2017/343 10.1145/3331184.3331267 10.1145/2365952.2365989 10.1145/3626772.3657902 10.1145/3485447.3512166 10.1145/3336191.3371832 10.1145/3041021.3054207 10.1145/3154793 10.1109/TKDE.2019.2941938 10.1109/TKDE.2019.2946247 10.1007/978-3-030-15719-7_3 10.1145/3632751 10.1109/TKDE.2019.2924656 10.1016/j.dss.2020.113489 10.1007/978-981-97-5555-4_3 10.1145/3522762 10.1145/3460231.3474265 10.1145/1401890.1401969 10.1145/3616855.3635816 10.1145/3539618.3591723 10.1145/3543507.3583402 10.1145/3580305.3599768 10.1145/3442381.3449813 10.24963/ijcai.2019/587 10.1007/11760146_14 10.1145/3269206.3271684 10.1145/3663364 10.1007/s10489-024-05360-x 10.1145/3340531.3412012 10.1109/TKDE.2021.3104873 10.1007/978-3-540-73078-1_44 10.1145/3357384.3357992 10.1145/3488560.3498436 10.1145/3178876.3186150 10.1145/3539618.3591752 10.1145/2594455 10.1145/3397271.3401063 10.1609/aaai.v35i5.16515 10.1109/TKDE.2005.99 10.1145/2792838.2800173 10.1145/3336191.3371793
ContentType	Journal Article
DBID	97E RIA RIE AAYXX CITATION
DOI	10.1109/TKDE.2025.3565793
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005-present IEEE All-Society Periodicals Package (ASPP) 1998-Present IEEE Electronic Library (IEL) CrossRef
DatabaseTitle	CrossRef
DatabaseTitleList
Database_xml	– sequence: 1 dbid: RIE name: IEEE/IET Electronic Library url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering Computer Science
EISSN	1558-2191
EndPage	3932
ExternalDocumentID	10_1109_TKDE_2025_3565793 10980364
Genre	orig-research
GrantInformation_xml	– fundername: National Natural Science Foundation of China grantid: 62461160311; 62272315 funderid: 10.13039/501100001809 – fundername: National Key Research and Development Program of China grantid: 2023YFF0725100 – fundername: Basic Research Fund in Shenzhen Natural Science Foundation grantid: JCYJ20240813141441054
GroupedDBID	-~X .DC 0R~ 1OL 29I 4.4 5GY 5VS 6IK 97E 9M8 AAJGR AARMG AASAJ AAWTH ABAZT ABFSI ABQJQ ABVLG ACGFO ACIWK AENEX AETIX AGQYO AGSQL AHBIQ AI. AIBXA AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 E.L EBS EJD F5P HZ~ H~9 ICLAB IEDLZ IFIPE IFJZH IPLJI JAVBF LAI M43 MS~ O9- OCL P2P PQQKQ RIA RIE RNI RNS RXW RZB TAE TAF TN5 UHB VH1 AAYXX CITATION
ID	FETCH-LOGICAL-c220t-8475d6e7174d667c527c757859952637e4443108dbf18d88e2bc062ebc0909413
IEDL.DBID	RIE
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001504151700035&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	1041-4347
IngestDate	Sat Nov 29 07:48:32 EST 2025 Wed Aug 27 01:52:25 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Issue	7
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c220t-8475d6e7174d667c527c757859952637e4443108dbf18d88e2bc062ebc0909413
ORCID	0000-0001-9310-3460 0009-0001-3005-8225 0000-0001-5059-8360 0000-0001-9065-6852 0000-0001-6326-9531 0000-0003-3955-1195
PageCount	17
ParticipantIDs	crossref_primary_10_1109_TKDE_2025_3565793 ieee_primary_10980364
PublicationCentury	2000
PublicationDate	2025-07-01
PublicationDateYYYYMMDD	2025-07-01
PublicationDate_xml	– month: 07 year: 2025 text: 2025-07-01 day: 01
PublicationDecade	2020
PublicationTitle	IEEE transactions on knowledge and data engineering
PublicationTitleAbbrev	TKDE
PublicationYear	2025
Publisher	IEEE
Publisher_xml	– name: IEEE
References	ref57 ref12 ref56 ref15 ref14 ref53 Glorot (ref65) ref52 ref11 ref55 ref10 ref54 ref17 ref16 ref19 ref18 Ahangama (ref26) 2019 ref51 ref50 ref46 ref45 Rezende (ref59) ref48 ref47 ref42 Kipf (ref37) 2016 ref41 Chen (ref35) 2021; 41 ref44 Kingma (ref64) 2014 ref43 ref49 ref8 ref7 ref9 ref4 ref3 ref6 ref5 Xiang (ref24) 2024 ref40 ref34 ref36 Yuan (ref60) ref30 ref33 ref32 ref2 ref1 ref39 ref38 Kingma (ref58) 2013 ref23 ref25 Fernández-Tobías (ref31) ref20 ref63 ref22 ref21 ref28 ref27 ref29 Gong (ref13) ref62 ref61
References_xml	– ident: ref62 doi: 10.1145/3038912.3052569 – year: 2019 ident: ref26 article-title: Latent user linking for collaborative cross domain recommendation – ident: ref38 doi: 10.1109/ICDM58522.2023.00053 – ident: ref47 doi: 10.1109/TKDE.2023.3295601 – ident: ref57 doi: 10.1007/s13278-020-0626-2 – ident: ref3 doi: 10.1145/2724720 – year: 2013 ident: ref58 article-title: Auto-encoding variational Bayes – start-page: 249 volume-title: Proc. 13th Int. Conf. Artif. Intell. Statist. ident: ref65 article-title: Understanding the difficulty of training deep feedforward neural networks – ident: ref61 doi: 10.48550/arXiv.1205.2618 – start-page: 11480 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref60 article-title: Tenrec: A large-scale multipurpose benchmark dataset for recommender systems – ident: ref12 doi: 10.1073/pnas.1218772110 – ident: ref25 doi: 10.1109/TKDE.2023.3253168 – ident: ref15 doi: 10.1016/j.dss.2015.03.008 – ident: ref48 doi: 10.1145/1526709.1526899 – ident: ref55 doi: 10.1145/3616855.3635830 – ident: ref63 doi: 10.1145/3637528.3671789 – ident: ref51 doi: 10.1145/3038912.3052695 – volume: 41 issue: 12 year: 2021 ident: ref35 article-title: Staged variational autoencoder for heterogeneous one-class collaborative filtering publication-title: J. Comput. Appl. – ident: ref22 doi: 10.24963/ijcai.2017/343 – start-page: 1278 volume-title: Proc. 31th Int. Conf. Mach. Learn. ident: ref59 article-title: Stochastic backpropagation and approximate inference in deep generative models – year: 2014 ident: ref64 article-title: Adam: A method for stochastic optimization – ident: ref42 doi: 10.1145/3331184.3331267 – ident: ref11 doi: 10.1145/2365952.2365989 – ident: ref28 doi: 10.1145/3626772.3657902 – ident: ref29 doi: 10.1145/3485447.3512166 – ident: ref54 doi: 10.1145/3336191.3371832 – ident: ref4 doi: 10.1145/3041021.3054207 – year: 2016 ident: ref37 article-title: Variational graph auto-encoders publication-title: arXiv:1611.07308 – ident: ref52 doi: 10.1145/3154793 – start-page: 34 volume-title: Proc. 1st Workshop New Trends Content-Based Recommender Syst. Co-Located 8th ACM Conf. Recommender Syst. ident: ref31 article-title: Exploiting social tags in matrix factorization models for cross-domain collaborative filtering – ident: ref17 doi: 10.1109/TKDE.2019.2941938 – ident: ref16 doi: 10.1109/TKDE.2019.2946247 – ident: ref32 doi: 10.1007/978-3-030-15719-7_3 – ident: ref41 doi: 10.1145/3632751 – ident: ref21 doi: 10.1109/TKDE.2019.2924656 – ident: ref20 doi: 10.1016/j.dss.2020.113489 – start-page: 979 volume-title: Proc. 25th USENIX Secur. Symp. ident: ref13 article-title: You are who you know and how you behave: Attribute inference attacks via users’ social friends and behaviors – ident: ref23 doi: 10.1007/978-981-97-5555-4_3 – ident: ref9 doi: 10.1145/3522762 – ident: ref40 doi: 10.1145/3460231.3474265 – ident: ref2 doi: 10.1145/1401890.1401969 – ident: ref19 doi: 10.1145/3616855.3635816 – ident: ref45 doi: 10.1145/3539618.3591723 – ident: ref10 doi: 10.1145/3543507.3583402 – ident: ref46 doi: 10.1145/3580305.3599768 – ident: ref53 doi: 10.1145/3442381.3449813 – ident: ref5 doi: 10.24963/ijcai.2019/587 – ident: ref49 doi: 10.1007/11760146_14 – ident: ref6 doi: 10.1145/3269206.3271684 – ident: ref33 doi: 10.1145/3663364 – ident: ref39 doi: 10.1007/s10489-024-05360-x – ident: ref27 doi: 10.1145/3340531.3412012 – ident: ref30 doi: 10.1109/TKDE.2021.3104873 – ident: ref1 doi: 10.1007/978-3-540-73078-1_44 – ident: ref7 doi: 10.1145/3357384.3357992 – ident: ref36 doi: 10.1145/3488560.3498436 – ident: ref34 doi: 10.1145/3178876.3186150 – ident: ref18 doi: 10.1145/3539618.3591752 – ident: ref50 doi: 10.1145/2594455 – ident: ref43 doi: 10.1145/3397271.3401063 – ident: ref44 doi: 10.1609/aaai.v35i5.16515 – year: 2024 ident: ref24 article-title: Coherence-guided preference disentanglement for cross-domain recommendations – ident: ref14 doi: 10.1109/TKDE.2005.99 – ident: ref56 doi: 10.1145/2792838.2800173 – ident: ref8 doi: 10.1145/3336191.3371793
SSID	ssj0008781
Score	2.4659576
Snippet	Cross-domain recommendation (CDR) aims to alleviate the data sparsity problem by leveraging the benefits of modeling two domains. However, existing research...
SourceID	crossref ieee
SourceType	Index Database Publisher
StartPage	3916
SubjectTerms	Accuracy attribute inference attacks Autoencoders Computational modeling cross-domain recommendation (CDR) Data models Graph convolutional networks Information leakage Knowledge transfer Privacy Protection recommendation system Recommender systems Variational autoencoders
Title	Camouflaged Variational Graph AutoEncoder Against Attribute Inference Attacks for Cross-Domain Recommendation
URI	https://ieeexplore.ieee.org/document/10980364
Volume	37
WOSCitedRecordID	wos001504151700035&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE/IET Electronic Library customDbUrl: eissn: 1558-2191 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0008781 issn: 1041-4347 databaseCode: RIE dateStart: 19890101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELagYoCBQimivOSBCSkljV0_xqgPQKCKoVTdoiS2USWaoJLy-7lLUlQGBpYosmzJuYt99_n83RFyI5yR0mntGc4SD2mzXsJM4skk5rFmvrMlPXr2LCcTNZ_rl5qsXnJhrLXl5TPbxdcylm_ydI1HZbDCtcK42S7ZlVJUZK2fbVfJsiIpwAsARYzLOoQJY-6mT8MRQMGg32UY5dPslxHaqqpSGpVx85_TOSKHtfdIw0rdx2THZi3S3FRmoPVCbZGDrTSDJ2Q5iAHgu3fYOgydATiuDwDpPWarpuG6yEcZcttXNHyLF-Aw0rCoKmFZ-rhhBGIbEvIpuLl0gN_lDfMldKcIYZcw16o8U5u8jkfTwYNXl1nw0iDwCw_sU98IC7iOGyFk2g9kilnuMRVZIJi0nIOX4SuTuJ4yStkgSX0RWHhqAIc9dkoaWZ7ZM0ItOBCJZc4pGfBUpNolcY_HjjmjuGKqQ243co8-qmwaUYlCfB2hkiJUUlQrqUPaKPOtjpW4z_9ovyD7OLy6S3tJGsVqba_IXvpVLD5X1-XP8g2Hyr2M
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT8MwDLZ4ScCBN-JNDpyQCl2SNclxGoNNjInDmHar2iZBSGxDo-P3Y7cdGgcOXKoqSqvUbmJ_cT4b4CryVilvTGClSAOizQapsGmg0kQmRoTeFfToQVf1eno4NM8VWb3gwjjnisNn7oZui1i-nWQz2irDGW40xc2WYbUuJQ9LutbPwqtVUZMUAQbCIiFVFcTEp277j3ctBIO8fiMozmfELzO0UFelMCv32_8c0A5sVf4ja5QK34UlN96D7XltBlZN1T3YXEg0uA-jZoIQ37_j4mHZAOFxtQXIHihfNWvM8klrTOz2KWu8Jm_oMrJGXtbCcqwz5wRSG1HyGTq6rEnfFdxNRtidEYgd4VjLAk0H8HLf6jfbQVVoIcg4D_MALVTdRg6RnbRRpLI6VxnluadkZDwSykmJfkaobepr2mrteJqFEXd4NQgPa-IQVsaTsTsC5tCFSJ3wXisusygzPk1qMvHCWy210MdwPZd7_FHm04gLHBKamJQUk5LiSknHcEAyX-hYivvkj_ZLWG_3n7pxt9N7PIUNelV5svYMVvLpzJ3DWvaVv31OL4of5xuQN8DT
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Camouflaged+Variational+Graph+AutoEncoder+Against+Attribute+Inference+Attacks+for+Cross-Domain+Recommendation&rft.jtitle=IEEE+transactions+on+knowledge+and+data+engineering&rft.au=Xiong%2C+Yudi&rft.au=Guo%2C+Yongxin&rft.au=Pan%2C+Weike&rft.au=Yang%2C+Qiang&rft.date=2025-07-01&rft.pub=IEEE&rft.issn=1041-4347&rft.volume=37&rft.issue=7&rft.spage=3916&rft.epage=3932&rft_id=info:doi/10.1109%2FTKDE.2025.3565793&rft.externalDocID=10980364
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1041-4347&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1041-4347&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1041-4347&client=summon