Model Stealing Attacks Against Inductive Graph Neural Networks

Many real-world data come in the form of graphs. Graph neural networks (GNNs), a new family of machine learning (ML) models, have been proposed to fully leverage graph data to build powerful applications. In particular, the inductive GNNs, which can generalize to unseen data, become mainstream in th...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Proceedings - IEEE Symposium on Security and Privacy s. 1175 - 1192
Hlavní autoři: Shen, Yun, He, Xinlei, Han, Yufei, Zhang, Yang
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.05.2022
Témata:
ISSN:2375-1207
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Many real-world data come in the form of graphs. Graph neural networks (GNNs), a new family of machine learning (ML) models, have been proposed to fully leverage graph data to build powerful applications. In particular, the inductive GNNs, which can generalize to unseen data, become mainstream in this direction. Machine learning models have shown great potential in various tasks and have been deployed in many real-world scenarios. To train a good model, a large amount of data as well as computational resources are needed, leading to valuable intellectual property. Previous research has shown that ML models are prone to model stealing attacks, which aim to steal the functionality of the target models. However, most of them focus on the models trained with images and texts. On the other hand, little attention has been paid to models trained with graph data, i.e., GNNs. In this paper, we fill the gap by proposing the first model stealing attacks against inductive GNNs. We systematically define the threat model and propose six attacks based on the adversary's background knowledge and the responses of the target models. Our evaluation on six benchmark datasets shows that the proposed model stealing attacks against GNNs achieve promising performance. 1
AbstractList Many real-world data come in the form of graphs. Graph neural networks (GNNs), a new family of machine learning (ML) models, have been proposed to fully leverage graph data to build powerful applications. In particular, the inductive GNNs, which can generalize to unseen data, become mainstream in this direction. Machine learning models have shown great potential in various tasks and have been deployed in many real-world scenarios. To train a good model, a large amount of data as well as computational resources are needed, leading to valuable intellectual property. Previous research has shown that ML models are prone to model stealing attacks, which aim to steal the functionality of the target models. However, most of them focus on the models trained with images and texts. On the other hand, little attention has been paid to models trained with graph data, i.e., GNNs. In this paper, we fill the gap by proposing the first model stealing attacks against inductive GNNs. We systematically define the threat model and propose six attacks based on the adversary's background knowledge and the responses of the target models. Our evaluation on six benchmark datasets shows that the proposed model stealing attacks against GNNs achieve promising performance. 1
Author Shen, Yun
Zhang, Yang
Han, Yufei
He, Xinlei
Author_xml – sequence: 1
  givenname: Yun
  surname: Shen
  fullname: Shen, Yun
  organization: Norton Research Group
– sequence: 2
  givenname: Xinlei
  surname: He
  fullname: He, Xinlei
  organization: CISPA Helmholtz Center for Information Security
– sequence: 3
  givenname: Yufei
  surname: Han
  fullname: Han, Yufei
  organization: INRIA
– sequence: 4
  givenname: Yang
  surname: Zhang
  fullname: Zhang, Yang
  organization: CISPA Helmholtz Center for Information Security
BookMark eNotj8tKw0AUQEdRsK39Al3kBxLnzns2Qii2FuoDqusyydzUsTEpmani31uwqwNnceCMyUXXd0jILdACgNq79atQDETBKGOFNZwrqs_I1GoDSkkBHJQ9JyPGtcyBUX1FxjF-Usoot2JE7p96j222Tuja0G2zMiVX72JWbl3oYsqWnT_UKXxjthjc_iN7xsPg2iPSTz_s4jW5bFwbcXrihLzPH95mj_nqZbGclas8AJiUO42-UsawCitqal0JadVRaM2M9NYplEJy1jCjG6ikV9ZKrwFqqYQHi3xCbv67ARE3-yF8ueF3c9rlf0u1SkU
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1109/SP46214.2022.9833607
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Computer Science
EISBN 9781665413169
1665413166
EISSN 2375-1207
EndPage 1192
ExternalDocumentID 9833607
Genre orig-research
GrantInformation_xml – fundername: Helmholtz Association
  funderid: 10.13039/501100009318
GroupedDBID 23M
29O
6IE
6IF
6IH
6IL
6IN
AAJGR
AAWTH
ABLEC
ACGFS
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IJVOP
M43
OCL
RIE
RIL
RIO
RNS
ID FETCH-LOGICAL-i118t-a7edb6882beb08c7b4596b6877285d9a6e54532f287f1b5d6995d711c564d19e3
IEDL.DBID RIE
IngestDate Wed Aug 27 02:37:20 EDT 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i118t-a7edb6882beb08c7b4596b6877285d9a6e54532f287f1b5d6995d711c564d19e3
PageCount 18
ParticipantIDs ieee_primary_9833607
PublicationCentury 2000
PublicationDate 2022-May
PublicationDateYYYYMMDD 2022-05-01
PublicationDate_xml – month: 05
  year: 2022
  text: 2022-May
PublicationDecade 2020
PublicationTitle Proceedings - IEEE Symposium on Security and Privacy
PublicationTitleAbbrev SP
PublicationYear 2022
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0020394
Score 2.52867
Snippet Many real-world data come in the form of graphs. Graph neural networks (GNNs), a new family of machine learning (ML) models, have been proposed to fully...
SourceID ieee
SourceType Publisher
StartPage 1175
SubjectTerms Benchmark testing
Computational modeling
Data models
Graph neural networks
Machine learning
Privacy
Reconstruction algorithms
Title Model Stealing Attacks Against Inductive Graph Neural Networks
URI https://ieeexplore.ieee.org/document/9833607
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEA61ePBUtRXf5ODRbTeb1-YiFLF6kFLwQW9lk8xKobSl3fr7zaRrRfDiKUvYbGDCZuZL5puPkBvvshQYZ4l1ygaAkvPEltonylgptGQeIkPu_VkPh_l4bEYNcrvjwgBATD6DLj7Gu3y_cBs8KuuZnHOF1PE9rfWWq7UDVyk3oqbGsdT0XkZCZQwPTbKsW4_7JaAS_ceg9b-ZD0nnh4hHRzsXc0QaMD8mrW8lBlr_mG1yh5JmM4rJuUgvp_2qQu487X8E4L-uKAp0xI2NPmKBaoolOYpZaGIO-LpD3gYPr_dPSa2MkEwDIKiSQoO3KgTHFmyaO22FNCp0hFA5l94UCkJgxLMywKGSWemVMdJrxpxUwjMD_IQ054s5nBLqjQmvCQjfUsKVaQ5FiApLg24rE7I4I200x2S5LX4xqS1x_nf3BTlAi28zAi9Js1pt4Irsu89qul5dxxX7Ata3lnI
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PS8MwFH6MKehp6ib-NgePdmvSJG0uwhDnxDkGTtltNMnrGIxNts6_36SrE8GLp5TQNPBC896XvO99ADfWsBBpRANtpHYAJYkCncU2kEoLHgtqsWDIvffifj8ZjdSgArdbLgwiFsln2PSPxV2-XZi1PyprqSSKpKeO7wjOGd2wtbbwKowUL8lxNFSt1wGXjPpjE8aa5chfEiqFB-nU_jf3ATR-qHhksHUyh1DB-RHUvrUYSPlr1uHOi5rNiE_P9QRz0s5zz54n7YmD_quceImOYmsjj75ENfFFOdKZa4os8FUD3joPw_tuUGojBFMHCfIgjdFq6cJjjTpMTKy5UNJ1uGA5EValEl1oFLHMAaKMamGlUsLGlBohuaUKo2OozhdzPAFilXKvcXTfktxkYYKpiwsz5R0X4yI9hbo3x_hjU_5iXFri7O_ua9jrDl96495T__kc9r31N_mBF1DNl2u8hF3zmU9Xy6ti9b4A3IuZuQ
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%3Abook&rft.genre=proceeding&rft.title=Proceedings+-+IEEE+Symposium+on+Security+and+Privacy&rft.atitle=Model+Stealing+Attacks+Against+Inductive+Graph+Neural+Networks&rft.au=Shen%2C+Yun&rft.au=He%2C+Xinlei&rft.au=Han%2C+Yufei&rft.au=Zhang%2C+Yang&rft.date=2022-05-01&rft.pub=IEEE&rft.eissn=2375-1207&rft.spage=1175&rft.epage=1192&rft_id=info:doi/10.1109%2FSP46214.2022.9833607&rft.externalDocID=9833607