GraphLIME: Local Interpretable Model Explanations for Graph Neural Networks
Recently, graph neural networks (GNN) were shown to be successful in effectively representing graph structured data because of their good performance and generalization ability. However, explaining the effectiveness of GNN models is a challenging task because of the complex nonlinear transformations...
Saved in:
| Published in | IEEE Transactions on Knowledge and Data Engineering Vol. 35; no. 7; pp. 6968 - 6972 |
|---|---|
| Main Authors | , , , , |
| Format | Journal Article |
| Language | English Japanese |
| Published |
New York
IEEE
01.07.2023
Institute of Electrical and Electronics Engineers (IEEE) The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Subjects | |
| Online Access | Get full text |
Cover
Loading…
| Abstract | Recently, graph neural networks (GNN) were shown to be successful in effectively representing graph structured data because of their good performance and generalization ability. However, explaining the effectiveness of GNN models is a challenging task because of the complex nonlinear transformations made over the iterations. In this paper, we propose GraphLIME, a local interpretable model explanation for graphs using the Hilbert-Schmidt Independence Criterion (HSIC) Lasso, which is a nonlinear feature selection method. GraphLIME is a generic GNN-model explanation framework that learns a nonlinear interpretable model locally in the subgraph of the node being explained. Through experiments on two real-world datasets, the explanations of GraphLIME are found to be of extraordinary degree and more descriptive in comparison to the existing explanation methods. |
|---|---|
| AbstractList | Recently, graph neural networks (GNN) were shown to be successful in effectively representing graph structured data because of their good performance and generalization ability. However, explaining the effectiveness of GNN models is a challenging task because of the complex nonlinear transformations made over the iterations. In this paper, we propose GraphLIME, a local interpretable model explanation for graphs using the Hilbert-Schmidt Independence Criterion (HSIC) Lasso, which is a nonlinear feature selection method. GraphLIME is a generic GNN-model explanation framework that learns a nonlinear interpretable model locally in the subgraph of the node being explained. Through experiments on two real-world datasets, the explanations of GraphLIME are found to be of extraordinary degree and more descriptive in comparison to the existing explanation methods. |
| Author | Singh, Dinesh Yamada, Makoto Tian, Yuan Huang, Qiang Chang, Yi |
| Author_xml | – sequence: 1 givenname: Qiang orcidid: 0000-0003-0046-0923 surname: Huang fullname: Huang, Qiang email: huangqiang18@mails.jlu.edu.cn organization: School of Artificial Intelligence, International Center of Future Science, Key Laboratory of Symbolic Computation and Knowledge Engineering, Jilin University, Changchun, Jilin, China – sequence: 2 givenname: Makoto orcidid: 0000-0001-7508-5094 surname: Yamada fullname: Yamada, Makoto email: makoto.yamada@riken.jp organization: Kyoto University, Kyoto, Japan – sequence: 3 givenname: Yuan surname: Tian fullname: Tian, Yuan email: yuantian@jlu.edu.cn organization: School of Artificial Intelligence, International Center of Future Science, Key Laboratory of Symbolic Computation and Knowledge Engineering, Jilin University, Changchun, Jilin, China – sequence: 4 givenname: Dinesh orcidid: 0000-0001-8889-9847 surname: Singh fullname: Singh, Dinesh email: dinesh.singh@riken.jp organization: RIKEN AIP, Kyoto, Japan – sequence: 5 givenname: Yi orcidid: 0000-0003-2697-8093 surname: Chang fullname: Chang, Yi email: yichang@jlu.edu.cn organization: School of Artificial Intelligence, International Center of Future Science, Key Laboratory of Symbolic Computation and Knowledge Engineering, Jilin University, Changchun, Jilin, China |
| BackLink | https://cir.nii.ac.jp/crid/1873961342498426240$$DView record in CiNii |
| BookMark | eNp9kD9PwzAQxS1UJNrCB0AskWBN8flfbDZUQqmalqXMkZO4IiXEwUkFfHuctmJgYLkb7v3u7r0RGtS2NghdAp4AYHW7XjzEE4IJmVCQEeP8BA0JJSKkUvABGgJmEDLKojM0atstxlhGEoZoMXO6eU3my_guSGyuq2Bed8Y1znQ6q0ywtIWpgvirqXStu9LWbbCxLthTwcrsnCdWpvu07q09R6cbXbXm4tjH6OUxXk-fwuR5Np_eJ2FOpexCoCwjUS6KwmQSYw4i46YgXKtISqE0L4CDUYJFQCRXjAJhmcxBYsCZKDQdo-vD3sbZj51pu3Rrd672J1MiCQVFqWBeBQdV7mzbOrNJG1e-a_edAk77zNI-s7TPLD1m5pnoD5OX3d5253RZ_UveHMi6LD3UVz-hSni3hCnJiCAMe9nVQVYaY34fUhKAgaA_r4qDIA |
| CODEN | ITKEEH |
| CitedBy_id | crossref_primary_10_3390_s23020634 crossref_primary_10_1016_j_ress_2024_110363 crossref_primary_10_1021_acs_chemrev_3c00189 crossref_primary_10_1080_07421222_2025_2452016 crossref_primary_10_1016_j_ipm_2023_103571 crossref_primary_10_1007_s00521_024_10868_x crossref_primary_10_1186_s13321_024_00820_5 crossref_primary_10_1016_j_inffus_2023_102190 crossref_primary_10_1093_schbul_sbac047 crossref_primary_10_1109_ACCESS_2024_3458050 crossref_primary_10_1109_JIOT_2023_3332848 crossref_primary_10_1038_s43586_024_00294_7 crossref_primary_10_1038_s41597_023_01974_x crossref_primary_10_1016_j_neucom_2024_127969 crossref_primary_10_1145_3696444 crossref_primary_10_1007_s10994_024_06576_1 crossref_primary_10_1007_s10489_024_06034_4 crossref_primary_10_1016_j_knosys_2024_112368 crossref_primary_10_1145_3711122 crossref_primary_10_1007_s41060_024_00636_4 crossref_primary_10_1109_TCE_2023_3303309 crossref_primary_10_1109_TSE_2024_3493245 crossref_primary_10_1145_3656464 crossref_primary_10_1145_3721129 crossref_primary_10_1016_j_apenergy_2024_124273 crossref_primary_10_3390_technologies12080128 crossref_primary_10_1109_ACCESS_2024_3440335 crossref_primary_10_1016_j_apenergy_2023_122151 crossref_primary_10_3390_computers14040119 crossref_primary_10_1007_s00521_025_11054_3 crossref_primary_10_1038_s41467_023_38192_3 crossref_primary_10_1007_s11633_024_1519_z crossref_primary_10_3389_frai_2023_1203546 crossref_primary_10_1109_TKDE_2024_3484461 crossref_primary_10_1109_TETCI_2024_3419714 crossref_primary_10_3389_fdata_2024_1392662 crossref_primary_10_1109_TNSM_2024_3436677 crossref_primary_10_3389_fpsyt_2023_1084443 crossref_primary_10_1002_aisy_202400304 crossref_primary_10_1016_j_ipm_2023_103600 crossref_primary_10_1007_s11042_023_17666_y crossref_primary_10_1016_j_ipm_2024_103821 crossref_primary_10_1145_3641543 crossref_primary_10_1016_j_ins_2023_119785 crossref_primary_10_1109_TAI_2024_3420262 crossref_primary_10_1021_acs_jcim_3c00396 crossref_primary_10_1021_acs_jcim_3c01642 crossref_primary_10_1109_TNSRE_2025_3530110 crossref_primary_10_1007_s13218_022_00781_7 crossref_primary_10_1016_j_neunet_2023_04_029 crossref_primary_10_1007_s10489_024_05546_3 crossref_primary_10_1109_TSIPN_2022_3180679 crossref_primary_10_1109_ACCESS_2022_3233036 crossref_primary_10_1038_s42256_025_00998_9 crossref_primary_10_1109_TPAMI_2022_3209686 crossref_primary_10_1007_s10489_023_05045_x crossref_primary_10_1134_S1064230724700138 crossref_primary_10_1145_3703922_3703930 crossref_primary_10_1016_j_osnem_2023_100247 crossref_primary_10_1109_OJCOMS_2024_3393853 crossref_primary_10_1109_TAI_2024_3459857 crossref_primary_10_1007_s11467_023_1325_z crossref_primary_10_1109_TPAMI_2022_3170302 crossref_primary_10_1016_j_patcog_2023_109991 crossref_primary_10_1016_j_iswa_2024_200353 crossref_primary_10_1109_TKDE_2024_3380709 crossref_primary_10_3390_rs16132504 crossref_primary_10_1080_13658816_2023_2254382 crossref_primary_10_1080_1206212X_2022_2133805 crossref_primary_10_1007_s11042_024_20271_2 crossref_primary_10_31857_S0002338824010122 crossref_primary_10_1002_cai2_136 crossref_primary_10_1111_risa_14241 crossref_primary_10_3390_info14080469 crossref_primary_10_1007_s10994_024_06733_6 crossref_primary_10_1109_ACCESS_2025_3537459 crossref_primary_10_1109_TNNLS_2022_3165618 crossref_primary_10_1016_j_neunet_2024_106885 |
| Cites_doi | 10.1109/TPAMI.2005.159 10.1007/978-3-030-86520-7_19 10.1162/NECO_a_00537 10.1007/978-3-319-46307-0_29 10.1145/3461702.3462562 10.1145/3219819.3219980 10.25300/MISQ/2014/38.1.04 10.1109/TSP.2020.3026980 10.1109/72.809084 10.1007/11564089_7 10.1109/TKDE.2018.2789451 10.1016/j.neucom.2016.12.038 10.1214/009053604000000067.MR2060166 10.1093/bioinformatics/btz333 10.18653/v1/N16-3020 |
| ContentType | Journal Article |
| Copyright | Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023 |
| Copyright_xml | – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023 |
| DBID | 97E RIA RIE RYH AAYXX CITATION 7SC 7SP 8FD JQ2 L7M L~C L~D |
| DOI | 10.1109/TKDE.2022.3187455 |
| DatabaseName | IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CiNii Complete CrossRef Computer and Information Systems Abstracts Electronics & Communications Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional |
| DatabaseTitle | CrossRef Technology Research Database Computer and Information Systems Abstracts – Academic Electronics & Communications Abstracts ProQuest Computer Science Collection Computer and Information Systems Abstracts Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Professional |
| DatabaseTitleList | Technology Research Database |
| Database_xml | – sequence: 1 dbid: RIE name: IEEE Electronic Library (IEL) url: https://proxy.k.utb.cz/login?url=https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Engineering Computer Science |
| EISSN | 2326-3865 1558-2191 |
| EndPage | 6972 |
| ExternalDocumentID | 10_1109_TKDE_2022_3187455 9811416 |
| Genre | orig-research |
| GrantInformation_xml | – fundername: Jilin Province Youth Science and Technology Program grantid: 20210508060RQ – fundername: National Natural Science Foundation of China grantid: U19A2065; 61976102 funderid: 10.13039/501100001809 – fundername: MEXT KAKENHI grantid: 20H04243 |
| GroupedDBID | -~X .DC 0R~ 29I 4.4 5GY 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACGFO ACIWK AENEX AGQYO AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD F5P HZ~ IEDLZ IFIPE IPLJI JAVBF LAI M43 MS~ O9- OCL P2P PQQKQ RIA RIE RNS RXW TAE TN5 UHB RYH AAYXX CITATION 7SC 7SP 8FD JQ2 L7M L~C L~D |
| ID | FETCH-LOGICAL-c388t-134b27c6ddeb800516b5ed25a978869a5d151e96471285943124b8c18010b6da3 |
| IEDL.DBID | RIE |
| ISSN | 1041-4347 |
| IngestDate | Mon Jun 30 02:35:54 EDT 2025 Thu Apr 24 23:01:23 EDT 2025 Tue Jul 01 01:19:41 EDT 2025 Fri Jun 27 00:23:08 EDT 2025 Wed Aug 27 02:04:47 EDT 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 7 |
| Language | English Japanese |
| 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-c388t-134b27c6ddeb800516b5ed25a978869a5d151e96471285943124b8c18010b6da3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| ORCID | 0000-0001-7508-5094 0000-0003-2697-8093 0000-0001-8889-9847 0000-0003-0046-0923 |
| PQID | 2823193364 |
| PQPubID | 85438 |
| PageCount | 5 |
| ParticipantIDs | crossref_citationtrail_10_1109_TKDE_2022_3187455 nii_cinii_1873961342498426240 crossref_primary_10_1109_TKDE_2022_3187455 ieee_primary_9811416 proquest_journals_2823193364 |
| ProviderPackageCode | CITATION AAYXX |
| PublicationCentury | 2000 |
| PublicationDate | 2023-07-01 |
| PublicationDateYYYYMMDD | 2023-07-01 |
| PublicationDate_xml | – month: 07 year: 2023 text: 2023-07-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | New York |
| PublicationPlace_xml | – name: New York |
| PublicationTitle | IEEE Transactions on Knowledge and Data Engineering |
| PublicationTitleAbbrev | TKDE |
| PublicationYear | 2023 |
| Publisher | IEEE Institute of Electrical and Electronics Engineers (IEEE) The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Publisher_xml | – name: IEEE – name: Institute of Electrical and Electronics Engineers (IEEE) – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| References | ref13 Yeh (ref8) ref12 Sundararajan (ref6) ref10 ref1 ref19 ref18 Koh (ref7) Hamilton (ref26) ref24 Lundberg (ref5) ref23 Chen (ref4) ref25 ref20 ref22 ref21 Yuan (ref17) Erhan (ref14) 2009 Shrikumar (ref15) ref28 ref27 Luo (ref16) 2020 Ying (ref9) Lakkaraju (ref2) 2017 ref3 Hu (ref11) 2021 |
| References_xml | – start-page: 9240 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref9 article-title: GNNExplainer: Generating explanations for graph neural networks – start-page: 1885 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref7 article-title: Understanding black-box predictions via influence functions – ident: ref24 doi: 10.1109/TPAMI.2005.159 – start-page: 4765 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref5 article-title: A unified approach to interpreting model predictions – start-page: 19 620 volume-title: Proc. Adv. Neural Inf. Process. Syst. year: 2020 ident: ref16 article-title: Parameterized explainer for graph neural network – ident: ref18 doi: 10.1007/978-3-030-86520-7_19 – start-page: 883 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref4 article-title: Learning to explain: An information-theoretic perspective on model interpretation – ident: ref20 doi: 10.1162/NECO_a_00537 – ident: ref12 doi: 10.1007/978-3-319-46307-0_29 – ident: ref19 doi: 10.1145/3461702.3462562 – ident: ref27 doi: 10.1145/3219819.3219980 – ident: ref28 doi: 10.25300/MISQ/2014/38.1.04 – ident: ref10 doi: 10.1109/TSP.2020.3026980 – year: 2021 ident: ref11 article-title: OGB-LSC: A large-scale challenge for machine learning on graphs publication-title: arXiv:2103.09430 – start-page: 9291 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref8 article-title: Representer point selection for explaining deep neural networks – ident: ref13 doi: 10.1109/72.809084 – start-page: 1024 volume-title: Proc. Adv. Neural Inf. Process. Syst. ident: ref26 article-title: Inductive representation learning on large graphs – ident: ref25 doi: 10.1007/11564089_7 – ident: ref21 doi: 10.1109/TKDE.2018.2789451 – ident: ref1 doi: 10.1016/j.neucom.2016.12.038 – year: 2009 ident: ref14 article-title: Visualizing higher-layer features of a deep network – start-page: 3145 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref15 article-title: Learning important features through propagating activation differences – ident: ref23 doi: 10.1214/009053604000000067.MR2060166 – ident: ref22 doi: 10.1093/bioinformatics/btz333 – start-page: 3319 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref6 article-title: Axiomatic attribution for deep networks – ident: ref3 doi: 10.18653/v1/N16-3020 – start-page: 12 241 volume-title: Proc. Int. Conf. Mach. Learn. ident: ref17 article-title: On explainability of graph neural networks via subgraph explorations – year: 2017 ident: ref2 article-title: Interpretable & explorable approximations of black box models publication-title: arXiv:1707.01154 |
| SSID | ssj0008781 ssib012349108 ssib016643522 ssib030785969 ssib000200689 ssib008799116 ssib004836754 ssib001015853 |
| Score | 2.6599038 |
| Snippet | Recently, graph neural networks (GNN) were shown to be successful in effectively representing graph structured data because of their good performance and... |
| SourceID | proquest crossref nii ieee |
| SourceType | Aggregation Database Enrichment Source Index Database Publisher |
| StartPage | 6968 |
| SubjectTerms | Computational modeling Data models explanation Feature extraction Graph neural networks Graph theory Graphical representations interpretability Kernel Mathematical models Predictive models Structured data Toy manufacturing industry |
| Title | GraphLIME: Local Interpretable Model Explanations for Graph Neural Networks |
| URI | https://ieeexplore.ieee.org/document/9811416 https://cir.nii.ac.jp/crid/1873961342498426240 https://www.proquest.com/docview/2823193364 |
| Volume | 35 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV07T8MwED5RJhgoT1FokQcmRNskthObDUEBUcrUSt2i2HEkRNUi2i78eu7yKE8hliiDL4ly9vk--_x9AKekY8UDi7CERw4BSuq3FUa9tpTGZr7ysiSjA86Dx_BuJO7HcrwG56uzMM65vPjMdeg238tPZ3ZJS2VdrTB798Ma1BC4FWe1VlFXRbkgKaILxERcROUOpu_p7rB_3UMkGAQIUIndXX6Zg3JRFZxZpk9PP-JxPsnc1GFQfV5RW_LcWS5Mx759Y2787_dvw1aZbbLLonvswJqb7kK9UnJg5cDehc1PtIR70L8lFmsMa70L9kBzHfsoTTQTx0g_bcKoei8p1hLnDFNfllsxYvtAi8eivHy-D6Ob3vDqrl2KLrQtV4qk6YUJIhti2DOKhmxopEsDmSDcVKFOZIo5gqPzqz5x32H-EQijLLrV90yYJvwA1qezqTsEZrSQPEl0gPYiQ09kkXCRli5SViaRboBXuSG2JSM5CWNM4hyZeDomz8Xkubj0XAPOViYvBR3HX433yAerhuXvb0ALfY1vpCu25BpTGoEwVBE7v_Aa0Kx6QVwO6Hkc0Hap5jwUR78_9Rg2SIm-qORtwvridelamK8szEneUd8BY6ngTg |
| linkProvider | IEEE |
| linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV1Lj9MwEB51y4HlQKEFbaEFHzgh2iZ-JDa31dJS6OPUSr1FseNIFVW6ou2FX89MkpanEJcoBzuvsWfmi8ffB_CGdKwEdwhLROwRoGThQKPXGyhlXR7qIE9z2uC8WEbTtfy8UZsGvLvshfHel8Vnfkin5Vp-tncn-lU2Mhqz9zC6ggcY9xWvdmtd_K6OS0lSxBeIioSM6zXMMDCj1ezDGLEg5whRid9d_RKFSlkVjC3FdvuHRy7DzKQFi_MDVtUlX4anox26b79xN_7vGzyBx3W-yW6rAfIUGr5oQ-us5cDqqd2GRz8RE3Zg9pF4rNGxjd-zOUU79qM40e48IwW1HaP6vbT6m3hgmPyyshcjvg_ssawKzA_PYD0Zr-6mg1p2YeCE1iROLy2PXYSOz2qatJFVPuMqRcCpI5OqDLMETztYQ2K_wwyES6sdGjYMbJSl4jk0i33hb4BZI5VIU8Oxv8zREnksfWyUj7VTaWy6EJzNkLiak5ykMXZJiU0Ck5DlErJcUluuC28vXe4rQo5_Ne6QDS4N68_fhT7aGu9IR2wpDCY1EoGoJn5-GXShdx4FST2lDwmnBVMjRCRf_P2qr-HhdLWYJ_NPy9lLuCZd-qqutwfN49eT72P2crSvykH7HUgO45g |
| 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=GraphLIME%3A+Local+Interpretable+Model+Explanations+for+Graph+Neural+Networks&rft.jtitle=IEEE+transactions+on+knowledge+and+data+engineering&rft.au=Huang%2C+Qiang&rft.au=Yamada%2C+Makoto&rft.au=Tian%2C+Yuan&rft.au=Singh%2C+Dinesh&rft.date=2023-07-01&rft.pub=The+Institute+of+Electrical+and+Electronics+Engineers%2C+Inc.+%28IEEE%29&rft.issn=1041-4347&rft.eissn=1558-2191&rft.volume=35&rft.issue=7&rft.spage=6968&rft_id=info:doi/10.1109%2FTKDE.2022.3187455&rft.externalDBID=NO_FULL_TEXT |
| 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 |