Trustworthy Graph Neural Networks: Aspects, Methods, and Trends
Open accessThe DOI matches its Crossref record; 52 of the 52 references it cites were found in a registry or an index.
Abstract
Graph neural networks (GNNs) have emerged as a series of competent graph learning methods for diverse real-world scenarios, ranging from daily applications such as recommendation systems and question answering to cutting-edge technologies such as drug discovery in life sciences and n-body simulation in astrophysics. However, task performance is not the only requirement for GNNs. Performance-oriented GNNs have exhibited potential adverse effects, such as vulnerability to adversarial attacks, unexplainable discrimination against disadvantaged groups, or excessive resource consumption in edge computing environments. To avoid these unintentional harms, it is necessary to build competent GNNs characterized by trustworthiness. To this end, we propose a comprehensive roadmap to build trustworthy GNNs from the view of the various computing technologies involved. In this survey, we introduce basic concepts and comprehensively summarize existing efforts for trustworthy GNNs from six aspects, including robustness, explainability, privacy, fairness, accountability, and environmental well-being. In addition, we highlight the intricate cross-aspect relations between the above six aspects of trustworthy GNNs. Finally, we present a thorough overview of trending directions for facilitating the research and industrialization of trustworthy GNNs.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
No free lunch theorems for optimization
Wolpert, D. H., & Macready, W. G. (1997) · IEEE Transactions on Evolutionary Computation · doi:10.1109/4235.585893 · Cited by 14,237 · cited 1×
“This conforms to the “no free lunch theory” [289].”
BackgroundA 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
Self-citationExplaining and Harnessing Adversarial Examples
Goodfellow, I. J., Shlens, J., & Szegedy, C. (2014) · arXiv · doi:10.48550/arxiv.1412.6572 · Cited by 10,381
A Theory of Justice: Revised Edition
Rawls, J. (1999) · doi:10.4159/9780674042582 · Cited by 9,455
Advances and Open Problems in Federated Learning
Kairouz, P., & McMahan, H. B. (2021) · Foundations and Trends in Machine Learning · doi:10.1561/2200000083 · Cited by 7,772
Network motifs: simple building blocks of complex networks.
Milo, R., Shen-Orr, S., Itzkovitz, S., Kashtan, N., Chklovskii, D., & Alon, U. (2002) · Science · doi:10.1126/science.298.5594.824 · Cited by 7,368 · cited 1×
“Sometimes, explanations of GNNs are presented in the form of subgraphs since subgraphs with specified patterns (i.e., motifs) are the building blocks of some more complex networks [153].”
Background“Why Should I Trust You?”: Explaining the Predictions of Any Classifier
Ribeiro, M., Singh, S., & Guestrin, C. (2016) · Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Demonstrations · doi:10.18653/v1/n16-3020 · Cited by 6,682
Federated Machine Learning: Concept and Applications
Yang, Q., Liu, Y., Chen, T., & Tong, Y. (2019) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3298981 · Cited by 6,322
Federated learning: Challenges, methods, and future directions
Li, T., Sahu, A. K., Talwalkar, A., & Smith, V. (2020) · IEEE Signal Processing Magazine · doi:10.1109/msp.2020.2975749 · Cited by 6,210 · cited 1×
“In the context of FL, multiple rounds of communication among different individuals greatly increase the time cost of completing the model training [222], [262].”
BackgroundA Survey on Bias and Fairness in Machine Learning
Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2019) · arXiv · doi:10.48550/arxiv.1908.09635 · Cited by 5,812
A Survey on Bias and Fairness in Machine Learning
Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2022) · ACM Computing Surveys · doi:10.1145/3457607 · Cited by 5,812
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 · cited 1×
“…in computational pathology) are also used to measure explanations. b) Research differences: Unlike explanations for general AI (e.g., saliency maps [41] and subimages [40]), GNN explanations (e.g., nodes, edges, and subgraphs) [31] inherit the irregularity and discreteness of graph data, which go…”
BackgroundMembership Inference Attacks Against Machine Learning Models
Shokri, R., Stronati, M., Song, C., & Shmatikov, V. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.41 · Cited by 4,754 · cited 1×
“AI Continuous-valued Perturbations [37] Sub-images [38], Saliency Maps [39] Individual Samples [40] IID *[41] Ordinary Metrics [42] Model Scale [10], Energy-intensive Architectures [43]”
BackgroundLewis, J. D., & Weigert, A. (1985) · Social Forces · doi:10.2307/2578601 · Cited by 4,163 · cited 1×
“It is widely recognized that trust relates to people’s perceived ability to rely on others (e.g., persons or systems); trust, in fact, can be considered the foundation of social order [26].”
BackgroundGraph Convolutional Neural Networks for Web-Scale Recommender Systems
Ying, R., He, R., Chen, K., Eksombatchai, P., Hamilton, W. L., & Leskovec, J. (2018) · Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3219819.3219890 · Cited by 3,733
Energy and Policy Considerations for Deep Learning in NLP
Strubell, E., Ganesh, A., & McCallum, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1355 · Cited by 3,684 · cited 1×
“One study on energy consumption [330] shows that training a common NLP model produces the same amount of carbon dioxide as a human would produce in seven years.”
BackgroundDwork, C., Hardt, M., Pitassi, T., Reingold, O., & Zemel, R. (2012) · Proceedings of the 3rd Innovations in Theoretical Computer Science Conference · doi:10.1145/2090236.2090255 · Cited by 3,501
Learning Important Features Through Propagating Activation Differences
Shrikumar, A., Greenside, P., & Kundaje, A. (2017) · arXiv · doi:10.48550/arxiv.1704.02685 · Cited by 2,761
Convolutional Networks on Graphs for Learning Molecular Fingerprints
Duvenaud, D., Maclaurin, D., Aguilera-Iparraguirre, J., Gómez-Bombarelli, R., Hirzel, T., Aspuru-Guzik, A., & Adams, R. P. (2015) · arXiv · doi:10.48550/arxiv.1509.09292 · Cited by 2,628
Signature Verification Using A "Siamese" Time Delay Neural Network
BROMLEY, J., BENTZ, J. W., BOTTOU, L., GUYON, I., LECUN, Y., MOORE, C., SÄCKINGER, E., & SHAH, R. (1993) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001493000339 · Cited by 2,077
Federated Learning with Non-IID Data
Zhao, Y., Li, M., Lai, L., Suda, N., Civin, D., & Chandra, V. (2018) · arXiv · doi:10.48550/arxiv.1806.00582 · Cited by 2,077
SecureML: A System for Scalable Privacy-Preserving Machine Learning
Mohassel, P., & Zhang, Y. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.12 · Cited by 1,936 · cited 2×
“There are three main types of techniques for security computation, including trusted executive environment (TEE) [255], [256], homomorphic encryption (HE) [257], [258], and multiparty secure computation (MPC) [259], [260], [261].”
BackgroundMethodZhu, L., & Han, S. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-63076-8_2 · Cited by 1,906
Sanity Checks for Saliency Maps
Adebayo, J., Gilmer, J., Muelly, M., Goodfellow, I., Hardt, M., & Kim, B. (2018) · arXiv · doi:10.48550/arxiv.1810.03292 · Cited by 1,294
Algorithmic Decision Making and the Cost of Fairness
Corbett-Davies, S., Pierson, E., Feller, A., Goel, S., & Huq, A. (2017) · Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/3097983.3098095 · Cited by 1,149
Fairness definitions explained
Verma, S., & Rubin, J. (2018) · Proceedings of the International Workshop on Software Fairness · doi:10.1145/3194770.3194776 · Cited by 1,124
A Survey of Graph Neural Networks for Recommender Systems: Challenges, Methods, and Directions
Gao, C., Zheng, Y., Li, N., Li, Y., Qin, Y., Piao, J., Quan, Y., Chang, J., Jin, D., He, X., et al. (2023) · ACM Transactions on Recommender Systems · doi:10.1145/3568022 · Cited by 1,082
Kusner, M. J., Loftus, J. R., Russell, C., & Silva, R. (2017) · arXiv · doi:10.48550/arxiv.1703.06856 · Cited by 995 · cited 1×
“A typical definition is demographic parity [279], which requires that the predictions ˆ Y of GNNs satisfy where Pr is the probabilityand S indicates whether the sample comes from the protected (e.g., female) group [280].”
BackgroundOlah, C., Mordvintsev, A., & Schubert, L. (2017) · Distill · doi:10.23915/distill.00007 · Cited by 880
GNNExplainer: Generating Explanations for Graph Neural Networks
Ying, R., Bourgeois, D., You, J., Zitnik, M., & Leskovec, J. (2019) · arXiv · doi:10.48550/arxiv.1903.03894 · Cited by 850
Designing Energy-Efficient Convolutional Neural Networks Using Energy-Aware Pruning
Yang, T.-J., Chen, Y.-H., & Sze, V. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.643 · Cited by 732 · cited 2×
“…considerations into GNN training [337] will also promote the development of efficient GNNs. Adaptive architecture designs aimed at saving energy [45], [359] have been proposed in the context of CNN models; thus, it would also be worthwhile to analyze the energy consumption of different…”
BackgroundMethodA graph placement methodology for fast chip design
Mirhoseini, A., Goldie, A., Yazgan, M., Jiang, J. W., Songhori, E., Wang, S., Lee, Y.-J., Johnson, E., Pathak, O., Nova, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03544-w · Cited by 618
Trustworthy AI: From Principles to Practices
Li, B., Qi, P., Liu, B., Di, S., Liu, J., Pei, J., Yi, J., & Zhou, B. (2023) · ACM Computing Surveys · doi:10.1145/3555803 · Cited by 612 · cited 4×
“For example, trustworthy systems should be able to obtain knowledge from limited training data and achieve convincing competence (i.e., high accuracy) on unseen data [11].”
BackgroundInfluentialSelf-citationExplainability 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
Towards Poisoning of Deep Learning Algorithms with Back-gradient Optimization
Muñoz-González, L., Biggio, B., Demontis, A., Paudice, A., Wongrassamee, V., Lupu, E. C., & Roli, F. (2017) · Proceedings of the 10th ACM Workshop on Artificial Intelligence and Security · doi:10.1145/3128572.3140451 · Cited by 549 · cited 2×
“Evasion attacks can also be executed by this approach, altering the associated traffic graph of the malware and evading detection [130].”
BackgroundAccountability of AI Under the Law: The Role of Explanation
Doshi-Velez, F., Kortz, M., Budish, R., Bavitz, C., Gershman, S. J., O'Brien, D., Scott, K., Shieber, S., Waldo, J., Weinberger, D., et al. (2017) · doi:10.2139/ssrn.3064761 · Cited by 457 · cited 1×
“5) Accountability: Accountability is another aspect of trustworthy GNNs that consider the ability to determine whether a system is performing in accordance with procedural and substantive standards [70].”
BackgroundCausal Intervention for Leveraging Popularity Bias in Recommendation
Zhang, Y., Feng, F., He, X., Wei, T., Song, C., Ling, G., & Zhang, Y. (2021) · Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval · doi:10.1145/3404835.3462875 · Cited by 443
GCAN: Graph-aware Co-Attention Networks for Explainable Fake News Detection on Social Media
Lu, Y.-J., & Li, C.-T. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.48 · Cited by 427
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
Wieringa, M. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3372833 · Cited by 321 · cited 2×
“…the responsibilities associated with individual roles or development steps in GNNs. a) Metrics: The accountability of GNNs can be measured by whether a standard evaluation process exists [36] and how comprehensive the specifications are established for each step or role of the GNN life cycle [44].”
BackgroundToward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims
Brundage, M., Avin, S., Wang, J., Belfield, H., Krueger, G., Hadfield, G., Khlaaf, H., Yang, J., Toner, H., Fong, R., et al. (2020) · arXiv · doi:10.48550/arxiv.2004.07213 · Cited by 306 · cited 2×
“There are a number of fields in which “trust” is defined and studied, including philosophy, psychology, sociology, economics, technology, and organizational management [10], [12], [13].”
BackgroundTrustworthy ai: A computational perspective
Liu, H., Wang, Y., Fan, W., Liu, X., Li, Y., Jain, S., Liu, Y., Jain, A., & Tang, J. (2023) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3546872 · Cited by 261
Anomaly Detection in Dynamic Graphs via Transformer
Liu, Y., Pan, S., Wang, Y. G., Xiong, F., Wang, L., Chen, Q., & Lee, V. C. (2023) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2021.3124061 · Cited by 169 · cited 1×
“Thus, it is essential to arrive at a reasonable definition of fairness for various kinds of graph tasks (e.g., exploring dynamic graphs [307] or heterophilic graphs [308]) and applications, which is essential for building fairness in trustworthy GNNs.”
BackgroundSelf-citationFairness Under Unawareness: Assessing Disparity When Protected Class Is Unobserved
Chen, J., Kallus, N., Mao, X., Svacha, G., & Udell, M. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287594 · Cited by 150
Coloring Molecules with Explainable Artificial Intelligence for Preclinical Relevance Assessment
Jiménez-Luna, J., Skalic, M., Weskamp, N., & Schneider, G. (2021) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.0c01344 · Cited by 96
Exact and Consistent Interpretation for Piecewise Linear Neural Networks: A Closed Form Solution
Chu, L., Hu, X., Hu, J., Wang, L., & Pei, J. (2018) · Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3219819.3220063 · Cited by 87
Self-citationFAHT: An Adaptive Fairness-aware Decision Tree Classifier
Zhang, W., & Ntoutsi, E. (2019) · Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2019/205 · Cited by 85
GNN-SubNet: disease subnetwork detection with explainable graph neural networks
Pfeifer, B., Saranti, A., & Holzinger, A. (2022) · Bioinformatics · doi:10.1093/bioinformatics/btac478 · Cited by 69
Histographs: graphs in histopathology
Anand, D., Gadiya, S., & Sethi, A. (2020) · Medical Imaging 2020: Digital Pathology · doi:10.1117/12.2550114 · Cited by 53 · cited 1×
“When GNNs are deployed in medical systems (e.g., breast cancer subtype classification [192] or histopathology analysis [193], [194]), prediction trust-worthiness can be improved by involving GNN explanation methods.”
Method(2020) · doi:10.14778/3494124.3494128 · Cited by 46 · cited 1×
“Remarks: Note that bias toward user groups with protected attributes (e.g., gender) should also be avoided [67] in other tasks such as financial analysis (e.g., loan default risk prediction [304], [305] and fraud detection [306]), recidivism predictions, and salary prediction [23].”
BackgroundBSQ: Exploring Bit-Level Sparsity for Mixed-Precision Neural Network Quantization
Yang, H., Duan, L., Chen, Y., & Li, H. (2021) · arXiv · doi:10.48550/arxiv.2102.10462 · Cited by 45
Impact Of Missing Data Imputation On The Fairness And Accuracy Of Graph Node Classifiers
Mansoor, H., Ali, S., Alam, S., Khan, M. A., Ul Hassan, U., & Khan, I. (2022) · 2022 IEEE International Conference on Big Data (Big Data) · doi:10.1109/bigdata55660.2022.10020694 · Cited by 6
Cited by
Unifying Large Language Models and Knowledge Graphs: A Roadmap
Pan, S., Luo, L., Wang, Y., Chen, C., Wang, J., & Wu, X. (2024) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2024.3352100 · Cited by 1,376
“To better unify LLMs and KGs, many advanced techniques need to be incorporated, such as multi-modal learning [261], graph neural network [262], and continuous learning [263].”
MethodSelf-citationJin, M., Koh, H. Y., Wen, Q., Zambon, D., Alippi, C., Webb, G. I., King, I., & Pan, S. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3443141 · Cited by 567
Self-citationA Survey of Graph Neural Networks for Social Recommender Systems
Sharma, K., Lee, Y.-C., Nambi, S., Salian, A., Shah, S., Kim, S.-W., & Kumar, S. (2024) · ACM Computing Surveys · doi:10.1145/3661821 · Cited by 268
Multimodal data integration for oncology in the era of deep neural networks: a review
Waqas, A., Tripathi, A., Ramachandran, R. P., Stewart, P. A., & Rasool, G. (2024) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2024.1408843 · Cited by 165
Metadata
- Authors
- Zhang, H., Wu, B., Yuan, X., Pan, S., Tong, H., & Pei, J.
- Venue
- Proceedings of the IEEE
- Published
- 2024-02
- Volume
- 112
- Issue
- 2
- Pages
- 97-139
- ISSN
- 0018-9219 (ISSN-L)
- 1558-2256
- OpenAlex
- W4393034172
- Cited by
- 144
- References
- 52
- FWCI
- 17.86
- Type
- journal article
- Language
- en
- Open access
- green
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.