Evaluating explainability for graph neural networks
Open accessAbstract
As explanations are increasingly used to understand the behavior of graph neural networks (GNNs), evaluating the quality and reliability of GNN explanations is crucial. However, assessing the quality of GNN explanations is challenging as existing graph datasets have no or unreliable ground-truth explanations. Here, we introduce a synthetic graph data generator, Shape GG en , which can generate a variety of benchmark datasets (e.g., varying graph sizes, degree distributions, homophilic vs. heterophilic graphs) accompanied by ground-truth explanations. The flexibility to generate diverse synthetic datasets and corresponding ground-truth explanations allows Shape GG en to mimic the data in various real-world areas. We include Shape GG en and several real-world graph datasets in a graph explainability library, G raph XAI. In addition to synthetic and real-world graph datasets with ground-truth explanations, G raph XAI provides data loaders, data processing functions, visualizers, GNN model implementations, and evaluation metrics to benchmark GNN explainability methods.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Focusing on the most predictive relationships (edges with importance scores >0.5, representing an average of 14.9% of edges from the KG), the model's performance showed a slight reduction from AUPRC = 0.890 (s.d. 0.006) to AUPRC = 0.886 (s.d. 0.005), indicating that TxGNN's Explainer effectively identifies key associations and explanations accurately reflect TxGNN's internal reasoning.
“Focusing on the most predictive relationships (that is, edges with importance scores >0.5, representing an average of 14.9% of edges from the KG), the model’s performance showed a slight reduction from AUPRC = 0.890 (s.d. 0.006) to AUPRC = 0.886 (s.d. 0.005), indicating that TxGNN’s Explainer effectively identifies key associations34 and explanations accurately reflect TxGNN’s internal reasoning.”
From A foundation model for clinician-centered drug repurposing (2024) · cites this paper as [34]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Matplotlib: a 2D graphics environment
Hunter, J. D. (2007) · Computing in Science & Engineering · doi:10.1109/mcse.2007.55 · Cited by 39,249 · see citing works
Emergence of Scaling in Random Networks
Barabási, A.-L., & Albert, R. (1999) · Science · doi:10.1126/science.286.5439.509 · Cited by 36,135 · see citing works
Statistical mechanics of complex networks
Albert, R., & Barabási, A.-L. (2002) · Reviews of Modern Physics · doi:10.1103/revmodphys.74.47 · Cited by 20,471 · see citing works
A Comprehensive Survey on Graph Neural Networks
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., & Yu, P. S. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.2978386 · Cited by 11,000 · see citing works
Exploring Network Structure, Dynamics, and Function Using NetworkX
Hagberg, A. A., Schult, D. A., & Swart, P. J. (2008) · Proceedings of the Python in Science Conference · doi:10.25080/tcwv9851 · Cited by 7,477 · see citing works
Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
Simonyan, K., Vedaldi, A., & Zisserman, A. (2013) · arXiv · doi:10.48550/arxiv.1312.6034 · Cited by 5,370 · see citing works
Collective Classification in Network Data
Sen, P., Namata, G., Bilgic, M., Getoor, L., Gallagher, B., & Eliassi‐Rad, T. (2008) · AI Magazine · doi:10.1609/aimag.v29i3.2157 · Cited by 3,293 · see citing works
ZINC 15 – Ligand Discovery for Everyone
Sterling, T., & Irwin, J. J. (2015) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5b00559 · Cited by 3,224 · see citing works
A Comprehensive Survey on Graph Neural Networks
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., & Yu, P. (2020) · UTS ePRESS (University of Technology Sydney) · Cited by 3,174 · see citing works
Metrics for evaluating 3D medical image segmentation: analysis, selection, and tool
Taha, A. A., & Hanbury, A. (2015) · BMC Medical Imaging · doi:10.1186/s12880-015-0068-x · Cited by 3,022 · see citing works
Modeling polypharmacy side effects with graph convolutional networks
Zitnik, M., Agrawal, M., & Leskovec, J. (2018) · Bioinformatics · doi:10.1093/bioinformatics/bty294 · Cited by 1,501 · see citing works
Self-citationOpen Graph Benchmark: Datasets for Machine Learning on GraphsPreprint
Hu, W., Fey, M., Žitnik, M., Dong, Y., Ren, H., Liu, B., Catasta, M., & Leskovec, J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.00687 · Cited by 1,439 · see citing works
Microsoft Academic Graph: When experts are not enough
Wang, K., Shen, Z., Huang, C., Wu, C.-H., Dong, Y.-X., & Kanakia, A. (2020) · Quantitative Science Studies · doi:10.1162/qss_a_00021 · Cited by 808 · see citing works
Derivation and Validation of Toxicophores for Mutagenicity Prediction
Kazius, J., McGuire, R., & Bursi, R. (2004) · Journal of Medicinal Chemistry · doi:10.1021/jm040835a · Cited by 668 · see citing works
Explainability in Graph Neural Networks: A Taxonomic Survey
Yuan, H., Yu, H., Gui, S., & Ji, S. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3204236 · Cited by 640 · see citing works
Explainability Methods for Graph Convolutional Neural Networks
Pope, P. E., Kolouri, S., Rostami, M., Martin, C. E., & Hoffmann, H. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.01103 · Cited by 553 · see citing works
Beyond Homophily in Graph Neural Networks: Current Limitations and Effective DesignsPreprint
Zhu, J., Yan, Y., Zhao, L., Heimann, M., Akoglu, L., & Koutra, D. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.11468 · Cited by 512 · see citing works
Deep Graph Library: Towards Efficient and Scalable Deep Learning on GraphsPreprint
Wang, M., Yu, L., Zheng, D., Gan, Q., Gai, Y. L., Ye, Z., Li, M., Zhou, J., Huang, Q., Ma, C., et al. (2019) · arXiv (Cornell University) · Cited by 490 · see citing works
GraphLIME: Local Interpretable Model Explanations for Graph Neural Networks
Huang, Q., Yamada, M., Tian, Y., Singh, D., & Chang, Y. (2022) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2022.3187455 · Cited by 455 · see citing works
GNNExplainer: Generating Explanations for Graph Neural Networks.
Ying, R., Bourgeois, D., You, J., Žitnik, M., & Leskovec, J. (2019) · PubMed · Cited by 409 · see citing works
Parameterized Explainer for Graph Neural Network
Luo, D., Cheng, W., Xu, D., Yu, W., Zong, B., Chen, H., & Zhang, X. (2020) · arXiv · doi:10.48550/arxiv.2011.04573 · Cited by 355 · see citing works
Node Similarity Preserving Graph Convolutional Networks
Jin, W., Derr, T., Wang, Y., Ma, Y., Liu, Z., & Tang, J. (2021) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3437963.3441735 · Cited by 269 · see citing works
A Gentle Introduction to Graph Neural Networks
Sánchez-Lengeling, B., Reif, E., Pearce, A., & Wiltschko, A. B. (2021) · Distill · doi:10.23915/distill.00033 · Cited by 247 · see citing works
Artificial intelligence foundation for therapeutic science
Huang, K., Fu, T., Gao, W., Zhao, Y., Roohani, Y., Leskovec, J., Coley, C. W., Xiao, C., Sun, J., & Zitnik, M. (2022) · Nature Chemical Biology · doi:10.1038/s41589-022-01131-2 · Cited by 207 · cited 2× · see citing works
“, Open Graph Benchmark (OGB) [17], GNNMark [18], GraphGT [19], MalNet [20], Graph Robustness Benchmark (GRB) [21], Therapeutics Data Commons [22, 23],”
Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and DevelopmentPreprint
Huang, K., Fu, T., Gao, W., Zhao, Y., Roohani, Y., Leskovec, J., Coley, C. W., Xiao, C., Sun, J., & Žitnik, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.09548 · Cited by 154 · see citing works
Network medicine framework for identifying drug-repurposing opportunities for COVID-19
Morselli Gysi, D., do Valle, Í., Zitnik, M., Ameli, A., Gan, X., Varol, O., Ghiassian, S. D., Patten, J. J., Davey, R. A., Loscalzo, J., et al. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2025581118 · Cited by 107 · see citing works
Self-citationModeling Polypharmacy Side Effects with Graph Convolutional NetworksPreprint
Žitnik, M., Agrawal, M., & Leskovec, J. (2018) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/258814 · Cited by 70 · see citing works
When Comparing to Ground Truth is Wrong
Faber, L., Moghaddam, A. K., & Wattenhofer, R. P. (2021) · doi:10.1145/3447548.3467283 · Cited by 41 · see citing works
GNNMark: A Benchmark Suite to Characterize Graph Neural Network Training on GPUs
Baruah, T., Shivdikar, K., Dong, S., Sun, Y., Mojumder, S. A., Jung, K., Abellán, J. L., Ukidave, Y., Joshi, A., Kim, J., et al. (2021) · IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS) · doi:10.1109/ispass51385.2021.00013 · Cited by 31 · see citing works
Graph Robustness Benchmark: Benchmarking the Adversarial Robustness of Graph Machine LearningPreprint
Zheng, Q., Zou, X., Dong, Y., Cen, Y., Yin, D., Xu, J., Yang, Y., & Tang, J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.04314 · Cited by 24 · see citing works
Probing GNN Explainers: A Rigorous Theoretical and Empirical Analysis of GNN Explanation MethodsPreprint
Agarwal, C., Žitnik, M., & Lakkaraju, H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.09078 · Cited by 19 · see citing works
Self-citationA Large-Scale Database for Graph Representation LearningPreprint
Freitas, S., Dong, Y., Neil, J., & Chau, D. H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.07682 · Cited by 17 · see citing works
GraphGT: Machine Learning Datasets for Graph Generation and Transformation
Du, Y., Wang, S., Guo, X., Cao, H., Hu, S., Jiang, J., Varala, A., Angirekula, A., & Zhao, L. (2021) · neural information processing systems · Cited by 13 · see citing works
Queen, O. (2022) · Harvard Dataverse · doi:10.7910/dvn/kulos8 · Cited by 2 · see citing works
Self-citation
Flagged first, then by citations
Research works printed but not linked: 17
- [1]Agarwal, C., Lakkaraju, H. & Zitnik, M. Towards a unified framework for fair and stable graph representation learning. In UAI (2021).
- [2]Sanchez-Lengeling et al. Evaluating attribution for graph neural networks. NeurIPS (2020).
- [3]Giunchiglia, V., Shukla, C. V., Gonzalez, G. & Agarwal, C. Towards training GNNs using explanation directed message passing. In The First Learning on Graphs Conference (2022).
- [6]Baldassarre, F. & Azizpour, H. Explainability techniques for graph convolutional networks. In ICML Workshop on LRGR (2019).
- [7]Faber, L. et al. Contrastive graph neural network explanation. In ICML Workshop on Graph Representation Learning and Beyond (2020).
- [9]Lucic, A., Ter Hoeve, M. A., Tolomei, G., De Rijke, M. & Silvestri, F. Cf-gnnexplainer: Counterfactual explanations for graph neural networks. In AISTATS (PMLR, 2022).
- [12]Schlichtkrull, M. S., De Cao, N. & Titov, I. Interpreting graph neural networks for nlp with differentiable edge masking. In ICLR (2021).
- [13]Vu, M. N. & Thai, M. T. PGM-Explainer: probabilistic graphical model explanations for graph neural networks. In NeurIPS (2020).
- [24]Wang, Z., Yin, H. & Song, Y. Benchmarking the combinatorial generalizability of complex query answering on knowledge graphs. In NeurIPS Datasets and Benchmarks (2021).
- [25]Liu, M. et al. DIG: A turnkey library for diving into graph deep learning research. JMLR (2021).
- [26]Fey, M. & Lenssen, J. E. Fast graph representation learning with PyTorch Geometric. ICLR 2019 (RLGM Workshop) (2019).
- [30]Sundararajan, M., Taly, A. & Yan, Q. Axiomatic attribution for deep networks. In ICML (2017).
- [31]Yuan, H., Yu, H., Wang, J., Li, K. & Ji, S. On explainability of graph neural networks via subgraph explorations. In ICML (2021).
- [32]Xu, K., Hu, W., Leskovec, J. & Jegelka, S. How powerful are graph neural networks? In ICLR (2019).
- [33]Kipf, T. N. & Welling, M. Semi-supervised classification with graph convolutional networks. In ICLR (2017).
- [41]Guyon, I. Design of experiments of the nips 2003 variable selection benchmark. In NIPS 2003 workshop on feature extraction and feature selection (2003).
- [42]Pedregosa, F. et al. Scikit-learn: Machine learning in Python. In JMLR (2011).
Cited by
Sorted by citations
A foundation model for clinician-centered drug repurposing
Huang, K., Chandak, P., Wang, Q., Havaldar, S., Vaid, A., Leskovec, J., Nadkarni, G. N., Glicksberg, B. S., Gehlenborg, N., & Zitnik, M. (2024) · Nature Medicine · doi:10.1038/s41591-024-03233-x · Cited by 249 · see citing works
“…an average of 14.9% of edges from the KG), the model’s performance showed a slight reduction from AUPRC = 0.890 (s.d. 0.006) to AUPRC = 0.886 (s.d. 0.005), indicating that TxGNN’s Explainer effectively identifies key associations 34 and explanations accurately reflect TxGNN’s internal reasoning.”
BackgroundMethodInfluentialSelf-citation
- Authors
- Agarwal, C., Queen, O., Lakkaraju, H., & Zitnik, M.
- Affiliations
- Harvard University (US)
- Venue
- Scientific Data
- Published
- 2023-03-18
- Volume
- 10
- Issue
- 1
- Pages
- 144-144
- ISSN
- 2052-4463 (ISSN-L)
- PMC
- PMC10024712
- OpenAlex
- W4327813569
- Cited by
- 186
- References
- 35
- Citations by year
- 2026: 60 · 2025: 70 · 2024: 40 · 2023: 14 · 2022: 2
- Type
- data paper
- Language
- en
- Open access
- gold
- doi.org
- Locations
- Scientific Data (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- arXiv (Cornell University) (cc-by)
- Funders
- National Science Foundation (2008461, FA8702-15-D-0001, 2030459, 2040989, 2033384)
- AstraZeneca
- U.S. Air Force (FA8702‐15‐D‐0001, Contract No. FA8702-15-D-0001, Air Force Contract No. FA8702-15-D-0001)
- Harvard Data Science Initiative, Harvard University
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Graph Neural Networks
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Keywords
- graph neural networks, explainability, explanation evaluation, benchmark dataset
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
