This is a collection of resources related to trustworthy graph neural networks.

• Related concepts
• Papers
  • Robustness
    • Attacks
    • Defences
  • Explainability
    • Interpretable GNNs
    • Post-hoc Explainers
  • Privacy
    • Privacy Attacks
    • Privacy-preserving Techniques for GNNs
  • Fairness
  • Accountability
  • Environmental well-being
    • Scalable GNN Architectures and Efficient Data Communication
    • Model Compression Methods
    • Efficient Frameworks and Accelerators
  • Others
  • Relations

1. Trustworthy Graph Neural Networks: Aspects, Methods and Trends. He Zhang, Bang Wu, Xingliang Yuan, Shirui Pan, Hanghang Tong, Jian Pei. Proceedings of the IEEE, 2024. paper
2. A Comprehensive Survey on Trustworthy Graph Neural Networks: Privacy, Robustness, Fairness, and Explainability. Enyan Dai, Tianxiang Zhao, Huaisheng Zhu, Junjie Xu, Zhimeng Guo, Hui Liu, Jiliang Tang, Suhang Wang. 2022. paper

1. A Comprehensive Survey on Graph Neural Networks. Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, Philip S. Yu. 2019. paper
2. Graph Neural Networks: Foundations, Frontiers, and Applications. Lingfei Wu, Peng Cui, Jian Pei, Liang Zhao. 2022. book

1. Trustworthy AI: A computational perspective. Haochen Liu, Yiqi Wang, Wenqi Fan, Xiaorui Liu, Yaxin Li, Shaili Jain, Yunhao Liu, Anil K. Jain, Jiliang Tang. 2021. paper
2. Trustworthy AI: from principles to practices. Bo Li], Peng Qi, Bo Liu, Shuai Di, Jingen Liu, Jiquan Pei, Jinfeng Yi, Bowen Zhou. 2021 paper
3. Trustworthy Machine Learning. Kush R. Varshney. 2022. book

Here we only list some papers. For other studies, please visit our Survey on Trustworthy GNNs.

1. Adversarial attack on graph structured data. ICML 2018. paper
2. Topology attack and defense for graph neural networks: An optimization perspective. IJCAI 2019. paper
3. Adversarial examples for graph data: Deep insights into attack and defense. IJCAI 2019. paper
4. Fast gradient attack on network embedding. Arxiv 2018. paper
5. Derivative-free optimization adversarial attacks for graph convolutional networks. PeerJ Computer Science 2021. paper
6. Adversarial attacks on graph neural networks via node injections: A hierarchical reinforcement learning approach. WWW 2020. paper
7. Adversarial attacks on neural networks for graph data. KDD 2018. paper

1. All you need is low (rank): Defending against adversarial attacks on graphs. WSDM 2020. paper
2. Graph structure learning for robust graph neural networks. KDD 2020. paper
3. Graph sanitation with application to node classification. WWW 2022. paper
4. Robust graph convolutional networks against adversarial attacks. KDD 2019. paper
5. Transferring robustness for graph neural network against poisoning attacks. WSDM 2020. paper
6. Defending graph convolutional networks against adversarial attacks. IEEE ICASSP 2020. paper
7. Gnnguard: Defending graph neural networks against adversarial attacks. NeurIPS 2020. paper
8. Graph adversarial training: Dynamically regularizing based on graph structure. IEEE TKDE 2021. paper
9. Robust training of graph convolutional networks via latent perturbation. PKDD 2020. paper
10. Topology attack and defense for graph neural networks: An optimization perspective. IJCAI 2019. paper
11. Certifiable robustness to graph perturbations. NeurIPS 2019. paper
12. Certifiable robustness and robust training for graph convolutional networks. KDD 2019. paper
13. Adversarial immunization for certifiable robustness on graphs. WSDM 2021. paper
14. Comparing and detecting adversarial attacks for graph deep learning. ICLR 2019. paper

1. Convolutional networks on graphs for learning molecular fingerprints. NeurIPS 2015. paper
2. Substructure assembling network for graph classification. AAAI 2018. paper
3. Towards explainable representation of time-evolving graphs via spatial-temporal graph attention networks. CIKM 2019. paper
4. Towards self-explainable graph neural network. CIKM 2021. paper
5. Protgnn: Towards self-explaining graph neural networks. AAAI 2022. paper
6. Motif-driven contrastive learning of graph representations. AAAI 2021. paper
7. Discovering invariant rationales for graph neural networks. ICLR 2022. paper
8. Graph information bottleneck for subgraph recognition. ICLR 2021. paper

1. Explainability techniques for graph convolutional networks. ICML 2019. paper
2. Explainability methods for graph convolutional neural networks. CVPR 2019. paper
3. Gnnexplainer: Generating explanations for graph neural networks. NeurIPS 2019. paper
4. Parameterized explainer for graph neural network. NeurIPS 2020. paper
5. Hard masking for explaining graph neural networks. OpenReview 2021. paper
6. Causal screening to interpret graph neural networks. OpenReview 2020. paper
7. Interpreting graph neural networks for NLP with differentiable edge masking. ICLR 2021. paper
8. On explainability of graph neural networks via subgraph explorations. ICML 2021. paper
9. Cf-gnnexplainer: Counterfactual explanations for graph neural networks. AISTATS 2022. paper
10. Robust counterfactual explanations on graph neural networks. NeurIPS 2021. paper
11. Towards multi-grained explainability for graph neural networks. NeurIPS 2021. paper
12. Learning and evaluating graph neural network explanations based on counterfactual and factual reasoning. WWW 2022. paper
13. Graphlime: Local interpretable model explanations for graph neural networks. Arxiv 2020. paper
14. Relex: A model-agnostic relational model explainer. AIES 2021. paper
15. Pgm-explainer: Probabilistic graphical model explanations for graph neural networks. NeurIPS 2020. paper
16. Higher-order explanations of graph neural networks via relevant walks. TPAMI 2021. paper
17. XGNN: towards model-level explanations of graph neural networks. KDD 2020. paper
18. Reinforcement learning enhanced explainer for graph neural networks. NeurIPS 2021. paper
19. Orphicx: A causality-inspired latent variable model for interpreting graph neural networks. CVPR 2022. paper
20. DEGREE: Decomposition based explanation for graph neural networks. ICLR 2021. paper
21. Counterfactual graphs for explainable classification of brain networks. KDD 2021. paper
22. Generative causal explanations for graph neural networks. ICML 2021. paper

1. Model extraction attacks on graph neural networks: Taxonomy and realization. AsiaCCS 2022. paper
2. Learning discrete structures for graph neural networks. ICML 2019. paper
3. Quantifying privacy leakage in graph embedding. MobiQuitous 2020. paper
4. Node-level membership inference attacks against graph neural networks. Arxiv 2021. paper
5. Stealing links from graph neural networks. USENIX Security Symposium 2021. paper
6. Adapting membership inference attacks to GNN for graph classification: Approaches and implications. IEEE ICDM 2021. paper
7. Membership inference attacks on knowledge graphs. Arxiv 2021. paper
8. Inference attacks against graph neural networks. USENIX Security Symposium 2022. paper
9. Graphmi: Extracting private graph data from graph neural networks. IJCAI 2021. paper
10. Linkteller: Recovering private edges from graph neural networks via influence analysis. IEEE S&P 2022. paper
11. Privacy-preserving representation learning on graphs: A mutual information perspective. KDD 2021. paper

1. Federated dynamic graph neural networks with secure aggregation for video-based distributedsurveillance. IEEE TIST 2022. paper
2. Spreadgnn: Serverless multi-task federated learning for graph neural networks. Arxiv 2021. paper
3. Federated graph classification over non-iid graphs. NeurIPS. paper
4. A federated multigraph integration approach for connectional brain template learning. ML-CDS 2021. paper
5. Federated learning of molecular properties in a heterogeneous setting. Arxiv 2021. paper
6. STFL: A temporal-spatial federated learning framework for graph neural networks. Arxiv 2021. paper
7. Fedgnn: Federated graph neural network for privacy-preserving recommendation. Arxiv 2021. paper
8. Federated social recommendation with graph neural network. Arxiv 2021. paper
9. A vertical federated learning framework for graph convolutional network. Arxiv 2021. paper
10. Vertically federated graph neural network for privacypreserving node classification. Arxiv 2020. paper
11. ASFGNN: automated separated-federated graph neural network. Peer-to-Peer Networking and Applications 2021. paper
12. Graphfl: A federated learning framework for semi-supervised node classification on graphs. Arxiv 2020. paper
13. Fedgl: Federated graph learning framework with global self-supervision. Arxiv 2021. paper
14. Cross-node federated graph neural network for spatio-temporal data modeling. KDD 2021. paper
15. Subgraph federated learning with missing neighbor generation. NeurIPS 2021. paper
16. Fedgraph: Federated graph learning with intelligent sampling. IEEE TPDS 2022. paper
17. Towards representation identical privacy-preserving graph neural network via split learning. Arxiv 2021. paper
18. Fedgraphnn: A federated learning system and benchmark for graph neural networks. Arxiv 2021. paper

1. Locally private graph neural networks. ACM CCS 2021. paper
2. Graph embedding for recommendation against attribute inference attacks. WWW 2021. paper

1. Netfense: Adversarial defenses against privacy attacks on neural networks for graph data. IEEE TKDE 2021. paper
2. Information obfuscation of graph neural networks. ICML 2021. paper
3. Adversarial privacypreserving graph embedding against inference attack. IEEE ITJ 2021. paper

1. Compositional fairness constraints for graph embeddings. ICML 2019. paper
2. Say no to the discrimination: Learning fair graph neural networks with limited sensitive attribute information. WSDM 2021. paper
3. Towards a unified framework for fair and stable graph representation learning. UAI 2021. paper
4. EDITS: modeling and mitigating data bias for graph neural networks. WWW 2022. paper
5. Inform: Individual fairness on graph mining. KDD 2020. paper
6. On dyadic fairness: Exploring and mitigating bias in graph connections. ICLR 2021. paper
7. Fairdrop: Biased edge dropout for enhancing fairness in graph representation learning. IEEE TAI 2021. paper
8. Individual fairness for graph neural networks: A ranking based approach. KDD 2021. paper

1. A pipeline for fair comparison of graph neural networks in node classification tasks. Arxiv 2020. paper
2. A novel genetic algorithm with hierarchical evaluation strategy for hyperparameter optimisation of graph neural networks. Arxiv 2021. paper
3. Bag of tricks of semi-supervised classification with graph neural networks. Arxiv 2021. paper
4. Bag of tricks for training deeper graph neural networks: A comprehensive benchmark study. IEEE TPAMI 2022. paper
5. A pipeline for fair comparison of graph neural networks in node classification tasks. Arxiv 2020. paper
6. A fair comparison of graph neural networks for graph classification. Arxiv 2019. paper
7. HASHTAG: hash signatures for online detection of fault-injection attacks on deep neural networks. ICCAD 2021. paper
8. Sensitive-sample fingerprinting of deep neural networks. CVPR 2019. paper
9. Proof-of-learning: Definitions and practice. IEEE S&P 2021. paper
10. Proof of learning (pole): Empowering machine learning with consensus building on blockchains (demo). AAAI 2021. paper

1. GraphSAINT: Graph Sampling Based Inductive Learning Method. ICLR 2020. paper
2. Gnnautoscale: Scalable and expressive graph neural networks via historical embeddings. ICML 2021. paper
3. Simplifying graph convolutional networks. ICML 2019. paper
4. Training graph neural networks with 1000 layers. ICML 2021. paper
5. Pinnersage: Multi-modal user embedding framework for recommendations at pinterest. KDD 2020. paper
6. ETA prediction with graph neural networks in google maps. CIKM 2021. paper
7. # Efficient Data Loader for Fast Sampling-Based GNN Training on Large Graphs. IEEE TPDS 2021. paper

1. On self-distilling graph neural network. IJCAI 2021. paper
2. Graph-free knowledge distillation for graph neural networks. IJCAI 2021. paper
3. Tinygnn: Learning efficient graph neural networks. KDD 2020. paper
4. A unified lottery ticket hypothesis for graph neural networks. ICML 2021. paper
5. Graph normalizing flows. NeurIPS 2019. paper
6. Binary graph neural networks. CVPR 2021. paper
7. Degree-quant: Quantization-aware training for graph neural networks. ICLR 2021. paper

1. Fast graph representation learning with PyTorch Geometric. ICLR 2019. paper
2. Deep graph library: Towards efficient and scalable deep learning on graphs. ICLR 2019. paper
3. Engn: A high-throughput and energy-efficient accelerator for large graph neural networks. IEEE TC 2021. paper
4. Hygcn: A GCN accelerator with hybrid architecture. HPCA 2020. paper
5. Characterizing and understanding gcns on GPU. IEEE CAL. paper
6. Alleviating irregularity in graph analytics acceleration: a hardware/software co-design approach. MICRO 2019. paper
7. Accelerating large scale real-time GNN inference using channel pruning. VLDB Endowment 2021. paper
8. G-cos: Gnnaccelerator co-search towards both better accuracy and efficiency. IEEE ICCAD. paper

1. How neural networks extrapolate: From feedforward to graph neural networks. ICLR 2021. paper

1. Explainability-based backdoor attacks against graph neural networks. WiseML 2021. paper
2. Jointly attacking graph neural network and its explanations. Arxiv 2021. paper
3. Towards a unified framework for fair and stable graph representation learning. UAI 2021. paper
4. Compositional fairness constraints for graph embeddings. ICML 2019. paper
5. Say no to the discrimination: Learning fair graph neural networks with limited sensitive attribute information. WSDM 2021. paper
6. Discrete-valued neural communication. NeurIPS 2021. paper
7. Graph structure learning for robust graph neural networks. KDD 2020. paper
8. Defensevgae: Defending against adversarial attacks on graph data via a variational graph autoencoder. Arxiv 2020. paper
9. Robust graph convolutional networks against adversarial attacks. KDD 2019. paper
10. Transferring robustness for graph neural network against poisoning attacks. WSDM 2020. paper
11. Extract the knowledge of graph neural networks and go beyond it: An effective knowledge distillation framework. WWW 2021. paper
12. Privacy-preserving representation learning on graphs: A mutual information perspective. KDD 2021. paper
13. Topological uncertainty: Monitoring trained neural networks through persistence of activation graphs. IJCAI 2021. paper

If you need more details, please visit the Survey on Trustworthy GNNs.

@article{DBLP:journals/corr/abs-2205-07424,
  author    = {He Zhang and
               Bang Wu and
               Xingliang Yuan and
               Shirui Pan and
               Hanghang Tong and
               Jian Pei},
  title     = {Trustworthy Graph Neural Networks: Aspects, Methods and Trends},
  journal   = {Proceedings of the IEEE},
  year      = {2024},
  doi       = {10.1109/JPROC.2024.3369017},
}