SPMGAE: Self-purified masked graph autoencoders release robust expression power
To tackle the scarcity of labeled graph data, graph self-supervised learning (SSL) has branched into two paradigms: Generative methods and Contrastive methods. Inspired by MAE and BERT in computer vision (CV) and natural language processing (NLP), masked graph autoencoders (MGAEs) are gaining popula...
Uloženo v:
| Vydáno v: | Neurocomputing (Amsterdam) Ročník 611; s. 128631 |
|---|---|
| Hlavní autoři: | , , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
Elsevier B.V
01.01.2025
|
| Témata: | |
| ISSN: | 0925-2312 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | To tackle the scarcity of labeled graph data, graph self-supervised learning (SSL) has branched into two paradigms: Generative methods and Contrastive methods. Inspired by MAE and BERT in computer vision (CV) and natural language processing (NLP), masked graph autoencoders (MGAEs) are gaining popularity in the generative genre. However, prevailing MGAEs are mostly designed under the assumption that the data has high homophilic score and is out of adversarial distortion. When people deliberately improve the performance on homophilic graph datasets, they ignore a critical issue that both internal heterophily and artificial attack noise are quite common in the real world. Therefore, when data itself is highly heterophilic or confronted with attacks, they merely have no defensive capability. Especially under self-supervised conditions, it is much more difficult to detect internal heterophily and resist artificial attacks. In this paper, we propose a Self-Purified Masked Graph Autoencoder (SPMGAE) to make up for the shortcomings of prevailing MGAEs in terms of robustness. SPMGAE first utilizes a self-purified module to prune raw graph data and separate perturbation information. The purified graph provides a robust graph structure for the entire pre-training process. Next, the encoding module reuses perturbation information for auxiliary training to enhance robustness, while the decoding module reconstructs the effective graph data at a finer granularity. Extensive experiments on homophilic and heterophilic datasets attacked by various attack methods demonstrate SPMGAE has a considerable robust expressive ability. Especially on small datasets with large perturbations, the improvement of defensive performance could reaches 10%–25%.
•We are the first to investigate the vulnerabilities of Mask Graph Autoencoders.•We analyze SOTA MGAEs to understand the reasons for their lack of robustness.•We propose a feasible plug-and-play self-purified scheme for MGAEs.•We propose a flexible robust pre-training network SPMGAE. |
|---|---|
| AbstractList | To tackle the scarcity of labeled graph data, graph self-supervised learning (SSL) has branched into two paradigms: Generative methods and Contrastive methods. Inspired by MAE and BERT in computer vision (CV) and natural language processing (NLP), masked graph autoencoders (MGAEs) are gaining popularity in the generative genre. However, prevailing MGAEs are mostly designed under the assumption that the data has high homophilic score and is out of adversarial distortion. When people deliberately improve the performance on homophilic graph datasets, they ignore a critical issue that both internal heterophily and artificial attack noise are quite common in the real world. Therefore, when data itself is highly heterophilic or confronted with attacks, they merely have no defensive capability. Especially under self-supervised conditions, it is much more difficult to detect internal heterophily and resist artificial attacks. In this paper, we propose a Self-Purified Masked Graph Autoencoder (SPMGAE) to make up for the shortcomings of prevailing MGAEs in terms of robustness. SPMGAE first utilizes a self-purified module to prune raw graph data and separate perturbation information. The purified graph provides a robust graph structure for the entire pre-training process. Next, the encoding module reuses perturbation information for auxiliary training to enhance robustness, while the decoding module reconstructs the effective graph data at a finer granularity. Extensive experiments on homophilic and heterophilic datasets attacked by various attack methods demonstrate SPMGAE has a considerable robust expressive ability. Especially on small datasets with large perturbations, the improvement of defensive performance could reaches 10%–25%.
•We are the first to investigate the vulnerabilities of Mask Graph Autoencoders.•We analyze SOTA MGAEs to understand the reasons for their lack of robustness.•We propose a feasible plug-and-play self-purified scheme for MGAEs.•We propose a flexible robust pre-training network SPMGAE. |
| ArticleNumber | 128631 |
| Author | Song, Shuhan Zhang, Yuan Cao, Huawei Ye, Xiaochun Li, Ping Dun, Ming |
| Author_xml | – sequence: 1 givenname: Shuhan surname: Song fullname: Song, Shuhan email: songshuhan19s@ict.ac.cn organization: State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China – sequence: 2 givenname: Ping surname: Li fullname: Li, Ping email: liping20b@ict.ac.cn organization: State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China – sequence: 3 givenname: Ming surname: Dun fullname: Dun, Ming email: dunming@ict.ac.cn organization: State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China – sequence: 4 givenname: Yuan surname: Zhang fullname: Zhang, Yuan email: zhangyuan-ams@ict.ac.cn organization: State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China – sequence: 5 givenname: Huawei orcidid: 0000-0003-1176-2521 surname: Cao fullname: Cao, Huawei email: caohuawei@ict.ac.cn organization: State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China – sequence: 6 givenname: Xiaochun orcidid: 0000-0003-4598-1685 surname: Ye fullname: Ye, Xiaochun email: yexiaochun@ict.ac.cn organization: State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China |
| BookMark | eNqFkL1OwzAUhT0UibbwBgx5gQRfO4mTDkhVVQpSUZEKs-U4N-CSxpGd8PP2pAoTA0xn-o7O-WZk0tgGCbkCGgGF9PoQNdhre4wYZXEELEs5TMiU5iwJGQd2TmbeHygFASyfkt3-8WGzXC-CPdZV2PbOVAbL4Kj82xAvTrWvgeo7i422JTofOKxReQycLXrfBfjZOvTe2CZo7Qe6C3JWqdrj5U_OyfPt-ml1F253m_vVchtqTtMu1CWlBRQKWRnHokpFgYlOBFIGIHgGnJaUcZWDTkVS5QKLWKcVhyTNdF4kwOdkMfZqZ713WEltOtUNOzqnTC2BypMOeZCjDnnSIUcdAxz_gltnjsp9_YfdjBgOx94NOum1GcRgaRzqTpbW_F3wDWkpgHI |
| CitedBy_id | crossref_primary_10_1038_s41598_025_08835_0 crossref_primary_10_1016_j_neucom_2025_129922 |
| Cites_doi | 10.1145/3580305.3599370 10.1145/3459637.3482092 10.1145/3539597.3570404 10.1145/3535101 10.1145/3219819.3220078 10.1145/3534678.3539484 10.1007/s10489-023-04836-6 10.1145/3394486.3403049 10.1109/TII.2020.2986316 10.1145/3132847.3132967 10.1145/3336191.3371789 10.1109/CVPR52688.2022.01553 10.1093/comnet/cnab014 10.1145/3308558.3313748 10.1145/3534678.3539321 10.1109/ICCV.2019.00662 10.1038/s41562-017-0290-3 10.1007/s10489-021-02587-w |
| ContentType | Journal Article |
| Copyright | 2024 Elsevier B.V. |
| Copyright_xml | – notice: 2024 Elsevier B.V. |
| DBID | AAYXX CITATION |
| DOI | 10.1016/j.neucom.2024.128631 |
| DatabaseName | CrossRef |
| DatabaseTitle | CrossRef |
| DatabaseTitleList | |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Computer Science |
| ExternalDocumentID | 10_1016_j_neucom_2024_128631 S0925231224014024 |
| GroupedDBID | --- --K --M .DC .~1 0R~ 123 1B1 1~. 1~5 4.4 457 4G. 53G 5VS 7-5 71M 8P~ 9JM 9JN AABNK AACTN AAEDT AAEDW AAIKJ AAKOC AALRI AAOAW AAQFI AAXKI AAXLA AAXUO AAYFN ABBOA ABCQJ ABFNM ABJNI ABMAC ACDAQ ACGFS ACRLP ACZNC ADBBV ADEZE AEBSH AEKER AENEX AFJKZ AFKWA AFTJW AFXIZ AGHFR AGUBO AGWIK AGYEJ AHHHB AHZHX AIALX AIEXJ AIKHN AITUG AJOXV AKRWK ALMA_UNASSIGNED_HOLDINGS AMFUW AMRAJ AOUOD AXJTR BKOJK BLXMC CS3 DU5 EBS EFJIC EO8 EO9 EP2 EP3 F5P FDB FIRID FNPLU FYGXN G-Q GBLVA GBOLZ IHE J1W KOM MO0 MOBAO N9A O-L O9- OAUVE OZT P-8 P-9 P2P PC. Q38 RIG ROL RPZ SDF SDG SDP SES SEW SPC SPCBC SSN SSV SSZ T5K ZMT ~G- 29N 9DU AAQXK AATTM AAYWO AAYXX ABWVN ABXDB ACLOT ACNNM ACRPL ACVFH ADCNI ADJOM ADMUD ADNMO AEIPS AEUPX AFPUW AGQPQ AIGII AIIUN AKBMS AKYEP ANKPU APXCP ASPBG AVWKF AZFZN CITATION EFKBS EFLBG EJD FEDTE FGOYB HLZ HVGLF HZ~ LG9 M41 R2- SBC WUQ XPP ~HD |
| ID | FETCH-LOGICAL-c306t-cd00b1bae2d447f67be5c57e0211738130d023a91c675f97eb4c6f31568c9b513 |
| ISICitedReferencesCount | 2 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001327273600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0925-2312 |
| IngestDate | Sat Nov 29 06:34:01 EST 2025 Tue Nov 18 22:00:30 EST 2025 Sat Nov 16 15:55:41 EST 2024 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Keywords | Masked graph autoencoders Graph neural networks Robustness Graph adversarial attacks |
| Language | English |
| LinkModel | OpenURL |
| MergedId | FETCHMERGED-LOGICAL-c306t-cd00b1bae2d447f67be5c57e0211738130d023a91c675f97eb4c6f31568c9b513 |
| ORCID | 0000-0003-1176-2521 0000-0003-4598-1685 |
| ParticipantIDs | crossref_citationtrail_10_1016_j_neucom_2024_128631 crossref_primary_10_1016_j_neucom_2024_128631 elsevier_sciencedirect_doi_10_1016_j_neucom_2024_128631 |
| PublicationCentury | 2000 |
| PublicationDate | 2025-01-01 2025-01-00 |
| PublicationDateYYYYMMDD | 2025-01-01 |
| PublicationDate_xml | – month: 01 year: 2025 text: 2025-01-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationTitle | Neurocomputing (Amsterdam) |
| PublicationYear | 2025 |
| Publisher | Elsevier B.V |
| Publisher_xml | – name: Elsevier B.V |
| References | Geisler, Schmidt, Şirin, Zügner, Bojchevski, Günnemann (b46) 2021; 34 Deng, Li, Feng, Zhang (b51) 2022 You, Chen, Wang, Shen (b37) 2020 Tu, Liao, Zhou, Peng, Ma, Liu, Liu, Cai (b14) 2023 Wu, Wang, Tyshetskiy, Docherty, Lu, Zhu (b49) 2019 Grill, Strub, Altché, Tallec, Richemond, Buchatskaya, Doersch, Avila Pires, Guo, Gheshlaghi Azar (b41) 2020; 33 N. Entezari, S.A. Al-Sayouri, A. Darvishzadeh, E.E. Papalexakis, All you need is low (rank) defending against adversarial attacks on graphs, in: Proceedings of the 13th International Conference on Web Search and Data Mining, 2020, pp. 169–177. Xiao, Zhu, Chen, Wang (b27) 2024; 36 Han, Karunasekera, Leckie (b3) 2020 Hassani, Khasahmadi (b23) 2020 Zhu, Yan, Zhao, Heimann, Akoglu, Koutra (b38) 2020; 33 K. He, X. Chen, S. Xie, Y. Li, P. Dollár, R. Girshick, Masked autoencoders are scalable vision learners, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 16000–16009. Li, Jin, Xu, Tang (b53) 2020 Hou, He, Cen, Liu, Dong, Kharlamov, Tang (b13) 2023 McInnes, Healy, Melville (b18) 2018 K. Li, Y. Liu, X. Ao, J. Chi, J. Feng, H. Yang, Q. He, Reliable representations make a stronger defender: Unsupervised structure refinement for robust gnn, in: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022, pp. 925–935. Jin, Derr, Liu, Wang, Wang, Liu, Tang (b36) 2020 Chen, Wu, Xu, Chen, Zheng, Xuan (b21) 2018 Yuan, Yu, Cao, Song, Xie, Wang (b31) 2023; 53 You, Chen, Sui, Chen, Wang, Shen (b24) 2020; 33 Guo, Wang (b4) 2020; 17 Kipf, Welling (b43) 2016 W. Jin, Y. Ma, X. Liu, X. Tang, S. Wang, J. Tang, Graph structure learning for robust graph neural networks, in: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2020, pp. 66–74. Z. Hou, X. Liu, Y. Cen, Y. Dong, H. Yang, C. Wang, J. Tang, Graphmae: Self-supervised masked graph autoencoders, in: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022, pp. 594–604. Zügner, Günnemann (b19) 2019 C. Wu, C. Wang, J. Xu, Z. Liu, K. Zheng, X. Wang, Y. Song, K. Gai, Graph Contrastive Learning with Generative Adversarial Network, in: Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2023, pp. 2721–2730. Waniek, Michalak, Wooldridge, Rahwan (b22) 2018; 2 K. Li, Y. Liu, X. Ao, Q. He, Revisiting graph adversarial attack and defense from a data distribution perspective, in: The Eleventh International Conference on Learning Representations, 2022. Xia, Wu, Wang, Chen, Li (b9) 2022 Garcia Duran, Niepert (b32) 2017; 30 Wang, Zhang, Zhu, Huang (b26) 2022 Wu, Sun, Zhang, Xie, Cui (b1) 2022; 55 Devlin, Chang, Lee, Toutanova (b10) 2018 Q. Tan, N. Liu, X. Huang, S.-H. Choi, L. Li, R. Chen, X. Hu, S2GAE: Self-Supervised Graph Autoencoders are Generalizable Learners with Graph Masking, in: Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, 2023, pp. 787–795. Suresh, Li, Hao, Neville (b30) 2021; 34 Pan, Hu, Long, Jiang, Yao, Zhang (b34) 2018 Velickovic, Cucurull, Casanova, Romero, Lio, Bengio (b44) 2017; 1050 Rozemberczki, Allen, Sarkar (b48) 2021; 9 Kipf, Welling (b17) 2016 D. Zügner, A. Akbarnejad, S. Günnemann, Adversarial attacks on neural networks for graph data, in: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2018, pp. 2847–2856. Thakoor, Tallec, Azar, Azabou, Dyer, Munos, Veličković, Valko (b25) 2021 Xu, Liu, Jiang, Xu, Mao, Chen, Zhou (b6) 2021 A. Chin, Y. Chen, K. M. Altenburger, J. Ugander, Decoupled smoothing on graphs, in: The World Wide Web Conference, 2019, pp. 263–272. C. Wang, S. Pan, G. Long, X. Zhu, J. Jiang, Mgae: Marginalized graph autoencoder for graph clustering, in: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management, 2017, pp. 889–898. J. Park, M. Lee, H.J. Chang, K. Lee, J.Y. Choi, Symmetric graph convolutional autoencoder for unsupervised graph representation learning, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 6519–6528. Li, Wu, Sun, Chen, Tian, Zhu, Meng, Zheng, Wang (b15) 2022 Zhang, Zitnik (b52) 2020; 33 Y. Wang, Z. Liu, Z. Fan, L. Sun, P.S. Yu, Dskreg: Differentiable sampling on knowledge graph for recommendation with relational gnn, in: Proceedings of the 30th ACM International Conference on Information & Knowledge Management, 2021, pp. 3513–3517. Bui, Cho, Yi (b5) 2022; 52 Yang, Zhang, Yang (b28) 2021; 34 Velickovic, Fedus, Hamilton, Liò, Bengio, Hjelm (b7) 2019; 2 Sen, Namata, Bilgic, Getoor, Galligher, Eliassi-Rad (b42) 2008; 29 E. Chien, J. Peng, P. Li, O. Milenkovic, Adaptive universal generalized pagerank graph neural network, 2020, arXiv preprint. Zhu, Xu, Yu, Liu, Wu, Wang (b8) 2020 McInnes (10.1016/j.neucom.2024.128631_b18) 2018 Xiao (10.1016/j.neucom.2024.128631_b27) 2024; 36 Zügner (10.1016/j.neucom.2024.128631_b19) 2019 Suresh (10.1016/j.neucom.2024.128631_b30) 2021; 34 Kipf (10.1016/j.neucom.2024.128631_b17) 2016 Hassani (10.1016/j.neucom.2024.128631_b23) 2020 Velickovic (10.1016/j.neucom.2024.128631_b44) 2017; 1050 10.1016/j.neucom.2024.128631_b29 Velickovic (10.1016/j.neucom.2024.128631_b7) 2019; 2 10.1016/j.neucom.2024.128631_b20 Waniek (10.1016/j.neucom.2024.128631_b22) 2018; 2 Zhang (10.1016/j.neucom.2024.128631_b52) 2020; 33 You (10.1016/j.neucom.2024.128631_b24) 2020; 33 Yang (10.1016/j.neucom.2024.128631_b28) 2021; 34 Garcia Duran (10.1016/j.neucom.2024.128631_b32) 2017; 30 Xu (10.1016/j.neucom.2024.128631_b6) 2021 Wu (10.1016/j.neucom.2024.128631_b1) 2022; 55 Wang (10.1016/j.neucom.2024.128631_b26) 2022 10.1016/j.neucom.2024.128631_b16 You (10.1016/j.neucom.2024.128631_b37) 2020 10.1016/j.neucom.2024.128631_b12 Chen (10.1016/j.neucom.2024.128631_b21) 2018 10.1016/j.neucom.2024.128631_b54 10.1016/j.neucom.2024.128631_b11 Rozemberczki (10.1016/j.neucom.2024.128631_b48) 2021; 9 10.1016/j.neucom.2024.128631_b50 10.1016/j.neucom.2024.128631_b2 Jin (10.1016/j.neucom.2024.128631_b36) 2020 Devlin (10.1016/j.neucom.2024.128631_b10) 2018 Thakoor (10.1016/j.neucom.2024.128631_b25) 2021 Deng (10.1016/j.neucom.2024.128631_b51) 2022 10.1016/j.neucom.2024.128631_b45 10.1016/j.neucom.2024.128631_b47 Han (10.1016/j.neucom.2024.128631_b3) 2020 Li (10.1016/j.neucom.2024.128631_b53) 2020 Kipf (10.1016/j.neucom.2024.128631_b43) 2016 10.1016/j.neucom.2024.128631_b40 Bui (10.1016/j.neucom.2024.128631_b5) 2022; 52 Geisler (10.1016/j.neucom.2024.128631_b46) 2021; 34 Grill (10.1016/j.neucom.2024.128631_b41) 2020; 33 Guo (10.1016/j.neucom.2024.128631_b4) 2020; 17 Wu (10.1016/j.neucom.2024.128631_b49) 2019 Pan (10.1016/j.neucom.2024.128631_b34) 2018 Li (10.1016/j.neucom.2024.128631_b15) 2022 Zhu (10.1016/j.neucom.2024.128631_b38) 2020; 33 Hou (10.1016/j.neucom.2024.128631_b13) 2023 Zhu (10.1016/j.neucom.2024.128631_b8) 2020 Sen (10.1016/j.neucom.2024.128631_b42) 2008; 29 10.1016/j.neucom.2024.128631_b39 Tu (10.1016/j.neucom.2024.128631_b14) 2023 10.1016/j.neucom.2024.128631_b35 Xia (10.1016/j.neucom.2024.128631_b9) 2022 Yuan (10.1016/j.neucom.2024.128631_b31) 2023; 53 10.1016/j.neucom.2024.128631_b33 |
| References_xml | – reference: N. Entezari, S.A. Al-Sayouri, A. Darvishzadeh, E.E. Papalexakis, All you need is low (rank) defending against adversarial attacks on graphs, in: Proceedings of the 13th International Conference on Web Search and Data Mining, 2020, pp. 169–177. – start-page: 4116 year: 2020 end-page: 4126 ident: b23 article-title: Contrastive multi-view representation learning on graphs publication-title: International Conference on Machine Learning – volume: 33 start-page: 5812 year: 2020 end-page: 5823 ident: b24 article-title: Graph contrastive learning with augmentations publication-title: Adv. Neural Inform. Process. Syst – volume: 9 start-page: cnab014 year: 2021 ident: b48 article-title: Multi-scale attributed node embedding publication-title: J. Complex Netw. – year: 2022 ident: b26 article-title: Augmentation-free graph contrastive learning with performance guarantee – volume: 36 year: 2024 ident: b27 article-title: Simple and asymmetric graph contrastive learning without augmentations publication-title: Adv. Neural Inf. Process. Syst. – volume: 33 start-page: 9263 year: 2020 end-page: 9275 ident: b52 article-title: Gnnguard: Defending graph neural networks against adversarial attacks publication-title: Adv. Neural Inf. Process. Syst. – year: 2020 ident: b53 article-title: Deeprobust: A pytorch library for adversarial attacks and defenses – reference: K. Li, Y. Liu, X. Ao, J. Chi, J. Feng, H. Yang, Q. He, Reliable representations make a stronger defender: Unsupervised structure refinement for robust gnn, in: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022, pp. 925–935. – year: 2021 ident: b25 article-title: Large-scale representation learning on graphs via bootstrapping – volume: 29 start-page: 93 year: 2008 ident: b42 article-title: Collective classification in network data publication-title: AI Mag. – volume: 55 start-page: 1 year: 2022 end-page: 37 ident: b1 article-title: Graph neural networks in recommender systems: a survey publication-title: ACM Comput. Surv. – volume: 1050 start-page: 10 year: 2017 end-page: 48550 ident: b44 article-title: Graph attention networks publication-title: Stat – reference: J. Park, M. Lee, H.J. Chang, K. Lee, J.Y. Choi, Symmetric graph convolutional autoencoder for unsupervised graph representation learning, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 6519–6528. – volume: 52 start-page: 2763 year: 2022 end-page: 2774 ident: b5 article-title: Spatial-temporal graph neural network for traffic forecasting: An overview and open research issues publication-title: Appl. Intell. – volume: 2 start-page: 139 year: 2018 end-page: 147 ident: b22 article-title: Hiding individuals and communities in a social network publication-title: Nat. Hum. Behav. – volume: 34 start-page: 15920 year: 2021 end-page: 15933 ident: b30 article-title: Adversarial graph augmentation to improve graph contrastive learning publication-title: Adv. Neural Inf. Process. Syst. – volume: 17 start-page: 2776 year: 2020 end-page: 2783 ident: b4 article-title: A deep graph neural network-based mechanism for social recommendations publication-title: IEEE Trans. Ind. Inform. – start-page: 10871 year: 2020 end-page: 10880 ident: b37 article-title: When does self-supervision help graph convolutional networks? publication-title: International Conference on Machine Learning – volume: 33 start-page: 21271 year: 2020 end-page: 21284 ident: b41 article-title: Bootstrap your own latent-a new approach to self-supervised learning publication-title: Adv. Neural Inf. Process. Syst. – reference: A. Chin, Y. Chen, K. M. Altenburger, J. Ugander, Decoupled smoothing on graphs, in: The World Wide Web Conference, 2019, pp. 263–272. – reference: Q. Tan, N. Liu, X. Huang, S.-H. Choi, L. Li, R. Chen, X. Hu, S2GAE: Self-Supervised Graph Autoencoders are Generalizable Learners with Graph Masking, in: Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, 2023, pp. 787–795. – year: 2022 ident: b15 article-title: Maskgae: masked graph modeling meets graph autoencoders – year: 2016 ident: b17 article-title: Variational graph auto-encoders – year: 2022 ident: b51 article-title: GARNET: Reduced-rank topology learning for robust and scalable graph neural networks publication-title: Learning on Graphs Conference – volume: 34 start-page: 7637 year: 2021 end-page: 7649 ident: b46 article-title: Robustness of graph neural networks at scale publication-title: Adv. Neural Inf. Process. Syst. – reference: E. Chien, J. Peng, P. Li, O. Milenkovic, Adaptive universal generalized pagerank graph neural network, 2020, arXiv preprint. – reference: Y. Wang, Z. Liu, Z. Fan, L. Sun, P.S. Yu, Dskreg: Differentiable sampling on knowledge graph for recommendation with relational gnn, in: Proceedings of the 30th ACM International Conference on Information & Knowledge Management, 2021, pp. 3513–3517. – reference: C. Wang, S. Pan, G. Long, X. Zhu, J. Jiang, Mgae: Marginalized graph autoencoder for graph clustering, in: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management, 2017, pp. 889–898. – year: 2019 ident: b49 article-title: Adversarial examples on graph data: Deep insights into attack and defense – reference: K. He, X. Chen, S. Xie, Y. Li, P. Dollár, R. Girshick, Masked autoencoders are scalable vision learners, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 16000–16009. – year: 2018 ident: b34 article-title: Adversarially regularized graph autoencoder for graph embedding – reference: C. Wu, C. Wang, J. Xu, Z. Liu, K. Zheng, X. Wang, Y. Song, K. Gai, Graph Contrastive Learning with Generative Adversarial Network, in: Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2023, pp. 2721–2730. – start-page: 24332 year: 2022 end-page: 24346 ident: b9 article-title: Progcl: Rethinking hard negative mining in graph contrastive learning publication-title: International Conference on Machine Learning – year: 2018 ident: b21 article-title: Fast gradient attack on network embedding – year: 2018 ident: b18 article-title: Umap: Uniform manifold approximation and projection for dimension reduction – year: 2020 ident: b8 article-title: Deep graph contrastive representation learning – volume: 53 start-page: 25154 year: 2023 end-page: 25170 ident: b31 article-title: Self-supervised robust graph neural networks against noisy graphs and noisy labels publication-title: Appl. Intell. – reference: D. Zügner, A. Akbarnejad, S. Günnemann, Adversarial attacks on neural networks for graph data, in: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2018, pp. 2847–2856. – year: 2018 ident: b10 article-title: Bert: Pre-training of deep bidirectional transformers for language understanding – year: 2016 ident: b43 article-title: Semi-supervised classification with graph convolutional networks – year: 2020 ident: b3 article-title: Graph neural networks with continual learning for fake news detection from social media – reference: Z. Hou, X. Liu, Y. Cen, Y. Dong, H. Yang, C. Wang, J. Tang, Graphmae: Self-supervised masked graph autoencoders, in: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022, pp. 594–604. – volume: 30 year: 2017 ident: b32 article-title: Learning graph representations with embedding propagation publication-title: Adv. Neural Inform Process. Syst – volume: 34 start-page: 14887 year: 2021 end-page: 14899 ident: b28 article-title: Graph adversarial self-supervised learning publication-title: Adv. Neural Inf. Process. Syst. – year: 2020 ident: b36 article-title: Self-supervised learning on graphs: Deep insights and new direction – reference: W. Jin, Y. Ma, X. Liu, X. Tang, S. Wang, J. Tang, Graph structure learning for robust graph neural networks, in: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2020, pp. 66–74. – year: 2019 ident: b19 article-title: Adversarial attacks on graph neural networks via meta learning publication-title: International Conference on Learning Representations – year: 2023 ident: b14 article-title: RARE: Robust masked graph autoencoder – year: 2021 ident: b6 article-title: MAF-GNN: Multi-adaptive spatiotemporal-flow graph neural network for traffic speed forecasting – volume: 2 start-page: 4 year: 2019 ident: b7 article-title: Deep graph infomax publication-title: ICLR (Poster) – reference: K. Li, Y. Liu, X. Ao, Q. He, Revisiting graph adversarial attack and defense from a data distribution perspective, in: The Eleventh International Conference on Learning Representations, 2022. – year: 2023 ident: b13 article-title: GraphMAE2: A decoding-enhanced masked self-supervised graph learner – volume: 33 start-page: 7793 year: 2020 end-page: 7804 ident: b38 article-title: Beyond homophily in graph neural networks: Current limitations and effective designs publication-title: Adv. Neural Inform. Process. Syst – ident: 10.1016/j.neucom.2024.128631_b29 doi: 10.1145/3580305.3599370 – year: 2020 ident: 10.1016/j.neucom.2024.128631_b53 – year: 2022 ident: 10.1016/j.neucom.2024.128631_b26 – ident: 10.1016/j.neucom.2024.128631_b45 – ident: 10.1016/j.neucom.2024.128631_b2 doi: 10.1145/3459637.3482092 – ident: 10.1016/j.neucom.2024.128631_b16 doi: 10.1145/3539597.3570404 – start-page: 24332 year: 2022 ident: 10.1016/j.neucom.2024.128631_b9 article-title: Progcl: Rethinking hard negative mining in graph contrastive learning – volume: 33 start-page: 9263 year: 2020 ident: 10.1016/j.neucom.2024.128631_b52 article-title: Gnnguard: Defending graph neural networks against adversarial attacks publication-title: Adv. Neural Inf. Process. Syst. – year: 2021 ident: 10.1016/j.neucom.2024.128631_b6 – year: 2018 ident: 10.1016/j.neucom.2024.128631_b10 – year: 2019 ident: 10.1016/j.neucom.2024.128631_b19 article-title: Adversarial attacks on graph neural networks via meta learning – volume: 34 start-page: 7637 year: 2021 ident: 10.1016/j.neucom.2024.128631_b46 article-title: Robustness of graph neural networks at scale publication-title: Adv. Neural Inf. Process. Syst. – year: 2018 ident: 10.1016/j.neucom.2024.128631_b34 – volume: 33 start-page: 7793 year: 2020 ident: 10.1016/j.neucom.2024.128631_b38 article-title: Beyond homophily in graph neural networks: Current limitations and effective designs publication-title: Adv. Neural Inform. Process. Syst – volume: 29 start-page: 93 issue: 3 year: 2008 ident: 10.1016/j.neucom.2024.128631_b42 article-title: Collective classification in network data publication-title: AI Mag. – year: 2023 ident: 10.1016/j.neucom.2024.128631_b14 – year: 2023 ident: 10.1016/j.neucom.2024.128631_b13 – volume: 55 start-page: 1 issue: 5 year: 2022 ident: 10.1016/j.neucom.2024.128631_b1 article-title: Graph neural networks in recommender systems: a survey publication-title: ACM Comput. Surv. doi: 10.1145/3535101 – ident: 10.1016/j.neucom.2024.128631_b20 doi: 10.1145/3219819.3220078 – ident: 10.1016/j.neucom.2024.128631_b40 doi: 10.1145/3534678.3539484 – volume: 30 year: 2017 ident: 10.1016/j.neucom.2024.128631_b32 article-title: Learning graph representations with embedding propagation publication-title: Adv. Neural Inform Process. Syst – year: 2020 ident: 10.1016/j.neucom.2024.128631_b36 – volume: 53 start-page: 25154 issue: 21 year: 2023 ident: 10.1016/j.neucom.2024.128631_b31 article-title: Self-supervised robust graph neural networks against noisy graphs and noisy labels publication-title: Appl. Intell. doi: 10.1007/s10489-023-04836-6 – volume: 34 start-page: 15920 year: 2021 ident: 10.1016/j.neucom.2024.128631_b30 article-title: Adversarial graph augmentation to improve graph contrastive learning publication-title: Adv. Neural Inf. Process. Syst. – year: 2016 ident: 10.1016/j.neucom.2024.128631_b43 – year: 2018 ident: 10.1016/j.neucom.2024.128631_b21 – year: 2016 ident: 10.1016/j.neucom.2024.128631_b17 – ident: 10.1016/j.neucom.2024.128631_b47 – volume: 33 start-page: 21271 year: 2020 ident: 10.1016/j.neucom.2024.128631_b41 article-title: Bootstrap your own latent-a new approach to self-supervised learning publication-title: Adv. Neural Inf. Process. Syst. – volume: 2 start-page: 4 issue: 3 year: 2019 ident: 10.1016/j.neucom.2024.128631_b7 article-title: Deep graph infomax publication-title: ICLR (Poster) – year: 2020 ident: 10.1016/j.neucom.2024.128631_b8 – ident: 10.1016/j.neucom.2024.128631_b54 doi: 10.1145/3394486.3403049 – volume: 17 start-page: 2776 issue: 4 year: 2020 ident: 10.1016/j.neucom.2024.128631_b4 article-title: A deep graph neural network-based mechanism for social recommendations publication-title: IEEE Trans. Ind. Inform. doi: 10.1109/TII.2020.2986316 – start-page: 4116 year: 2020 ident: 10.1016/j.neucom.2024.128631_b23 article-title: Contrastive multi-view representation learning on graphs – volume: 34 start-page: 14887 year: 2021 ident: 10.1016/j.neucom.2024.128631_b28 article-title: Graph adversarial self-supervised learning publication-title: Adv. Neural Inf. Process. Syst. – ident: 10.1016/j.neucom.2024.128631_b33 doi: 10.1145/3132847.3132967 – ident: 10.1016/j.neucom.2024.128631_b50 doi: 10.1145/3336191.3371789 – year: 2021 ident: 10.1016/j.neucom.2024.128631_b25 – start-page: 10871 year: 2020 ident: 10.1016/j.neucom.2024.128631_b37 article-title: When does self-supervision help graph convolutional networks? – volume: 36 year: 2024 ident: 10.1016/j.neucom.2024.128631_b27 article-title: Simple and asymmetric graph contrastive learning without augmentations publication-title: Adv. Neural Inf. Process. Syst. – ident: 10.1016/j.neucom.2024.128631_b11 doi: 10.1109/CVPR52688.2022.01553 – year: 2022 ident: 10.1016/j.neucom.2024.128631_b15 – year: 2020 ident: 10.1016/j.neucom.2024.128631_b3 – year: 2022 ident: 10.1016/j.neucom.2024.128631_b51 article-title: GARNET: Reduced-rank topology learning for robust and scalable graph neural networks – volume: 9 start-page: cnab014 issue: 2 year: 2021 ident: 10.1016/j.neucom.2024.128631_b48 article-title: Multi-scale attributed node embedding publication-title: J. Complex Netw. doi: 10.1093/comnet/cnab014 – ident: 10.1016/j.neucom.2024.128631_b39 doi: 10.1145/3308558.3313748 – ident: 10.1016/j.neucom.2024.128631_b12 doi: 10.1145/3534678.3539321 – volume: 1050 start-page: 10 issue: 20 year: 2017 ident: 10.1016/j.neucom.2024.128631_b44 article-title: Graph attention networks publication-title: Stat – year: 2018 ident: 10.1016/j.neucom.2024.128631_b18 – volume: 33 start-page: 5812 year: 2020 ident: 10.1016/j.neucom.2024.128631_b24 article-title: Graph contrastive learning with augmentations publication-title: Adv. Neural Inform. Process. Syst – ident: 10.1016/j.neucom.2024.128631_b35 doi: 10.1109/ICCV.2019.00662 – year: 2019 ident: 10.1016/j.neucom.2024.128631_b49 – volume: 2 start-page: 139 issue: 2 year: 2018 ident: 10.1016/j.neucom.2024.128631_b22 article-title: Hiding individuals and communities in a social network publication-title: Nat. Hum. Behav. doi: 10.1038/s41562-017-0290-3 – volume: 52 start-page: 2763 issue: 3 year: 2022 ident: 10.1016/j.neucom.2024.128631_b5 article-title: Spatial-temporal graph neural network for traffic forecasting: An overview and open research issues publication-title: Appl. Intell. doi: 10.1007/s10489-021-02587-w |
| SSID | ssj0017129 |
| Score | 2.4432647 |
| Snippet | To tackle the scarcity of labeled graph data, graph self-supervised learning (SSL) has branched into two paradigms: Generative methods and Contrastive methods.... |
| SourceID | crossref elsevier |
| SourceType | Enrichment Source Index Database Publisher |
| StartPage | 128631 |
| SubjectTerms | Graph adversarial attacks Graph neural networks Masked graph autoencoders Robustness |
| Title | SPMGAE: Self-purified masked graph autoencoders release robust expression power |
| URI | https://dx.doi.org/10.1016/j.neucom.2024.128631 |
| Volume | 611 |
| WOSCitedRecordID | wos001327273600001&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: PRVESC databaseName: Elsevier SD Freedom Collection Journals 2021 issn: 0925-2312 databaseCode: AIEXJ dateStart: 19950101 customDbUrl: isFulltext: true dateEnd: 99991231 titleUrlDefault: https://www.sciencedirect.com omitProxy: false ssIdentifier: ssj0017129 providerName: Elsevier |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1LT9wwELZa4NALbwQUkA_cVkHJOonj3lZoeVSCrrQgbU9R7NjimV3tJhU_v-NHstmCaDn0kkRW7ESZT-PPk_E3CB0D6VVCMOmxKONeSInwEkJyvYGZCC32QVRiik3Q6-tkNGID9wd_ZsoJ0KJIXl7Y5L-aGtrA2Hrr7AfM3QwKDXANRocjmB2O_2T44eDqvNfXK_2hfFLepJreK80zn7PZI5yMQnUnq8qxlrDUacymbgpMZp3pmFezUov-2-TYojPRNdTa_NVoeQhTCcLFGHrPWmoh17hqYgpDl-c7vKvuWjk_JnFgUM-Vlj3b1P15UxPA_lm5ni4k0Y1aIQkXW4Q2II4LbjZ2TtU6SpgWY-v-X_lwG054OClkpRN64AHhyfz2RcnsP6ayJsGwzl17SO0oqR4ltaN8RstdGjHw4su9y_7oe_PTiQZdK83o3r7eaWnSAV-_zdtMpsVObtbRqltW4J6Fwwb6JItNtFaX7MDOg2-hHxYd3_ACNrDFBjbYwG1sYIcNbLGB59jABhvb6Pasf3N64bmaGp6AxWHpidz3ecAz2c3DkKqYchmJiEqgegEF9kb8HFhcxgIBK0nFqOShiBWBVX4iGI8CsoOWinEhdxHmQSKzKFBhFLIQmB5XfpwomavApzz32R4i9fdJhROc13VPntL3rLOHvKbXxAqu_OV-Wn_61JFGSwZTwNO7Pfc_-KSv6Msc7AdoqZxW8hCtiF_l_Wx65MD0G_ksjdE |
| linkProvider | Elsevier |
| 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=SPMGAE%3A+Self-purified+masked+graph+autoencoders+release+robust+expression+power&rft.jtitle=Neurocomputing+%28Amsterdam%29&rft.au=Song%2C+Shuhan&rft.au=Li%2C+Ping&rft.au=Dun%2C+Ming&rft.au=Zhang%2C+Yuan&rft.date=2025-01-01&rft.issn=0925-2312&rft.volume=611&rft.spage=128631&rft_id=info:doi/10.1016%2Fj.neucom.2024.128631&rft.externalDBID=n%2Fa&rft.externalDocID=10_1016_j_neucom_2024_128631 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0925-2312&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0925-2312&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0925-2312&client=summon |