mlsanityData to 14 Oct 2026

Debiased Graph Neural Networks with Agnostic Label Selection Bias

arXiv 2201.07708 ↗First posted 19 Jan 2022cs.LG cs.AIcited by 25

Not on the map. Why?

Abstract (arXiv)

Most existing Graph Neural Networks (GNNs) are proposed without considering the selection bias in data, i.e., the inconsistent distribution between the training set with test set. In reality, the test data is not even available during the training process, making selection bias agnostic. Training GNNs with biased selected nodes leads to significant parameter estimation bias and greatly impacts the generalization ability on test nodes. In this paper, we first present an experimental investigation, which clearly shows that the selection bias drastically hinders the generalization ability of GNNs, and theoretically prove that the selection bias will cause the biased estimation on GNN parameters. Then to remove the bias in GNN estimation, we propose a novel Debiased Graph Neural Networks (DGNN) with a differentiated decorrelation regularizer. The differentiated decorrelation regularizer estimates a sample weight for each labeled node such that the spurious correlation of learned embeddings could be eliminated. We analyze the regularizer in causal view and it motivates us to differentiate the weights of the variables based on their contribution on the confounding bias. Then, these sample weights are used for reweighting GNNs to eliminate the estimation bias, thus help to improve the stability of prediction on unknown test nodes. Comprehensive experiments are conducted on several challenging graph datasets with two kinds of label selection biases. The results well verify that our proposed model outperforms the state-of-the-art methods and DGNN is a flexible framework to enhance existing GNNs.

Key facts. Cited by 25 papers in the map (data through 14 Oct 2026). Builds on 53 of its 105 references that are in the map. First posted on arXiv 19 Jan 2022.

Builds on

53 of 105 references.

References · 53

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  3. 2017Graph Attention Networkscited by 8,923
  4. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 7,005
  5. 2017Inductive Representation Learning on Large Graphscited by 6,320
  6. 2018How Powerful are Graph Neural Networks?cited by 4,393
  7. 2017Neural Message Passing for Quantum Chemistrycited by 3,487
  8. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  9. 2020Open Graph Benchmark: Datasets for Machine Learning on Graphscited by 2,163
  10. 2016Convolutional Neural Networks on Graphs with Fast Localized Spectral Filteringcited by 2,089
  11. 2019Invariant Risk Minimizationcited by 1,854
  12. 2017Modeling Relational Data with Graph Convolutional Networkscited by 1,694
38 more
  1. 2019Simplifying Graph Convolutional Networkscited by 1,477
  2. 2017MoleculeNet: A Benchmark for Molecular Machine Learningcited by 1,051
  3. 2018Predict then Propagate: Graph Neural Networks meet Personalized PageRankcited by 993
  4. 2016Revisiting Semi-Supervised Learning with Graph Embeddingscited by 976
  5. 2018Representation Learning on Graphs with Jumping Knowledge Networkscited by 877
  6. 2018Link Prediction Based on Graph Neural Networkscited by 832
  7. 2018Hierarchical Graph Representation Learning with Differentiable Poolingcited by 787
  8. 2020Out-of-Distribution Generalization via Risk Extrapolation (REx)cited by 742
  9. 2019GNNExplainer: Generating Explanations for Graph Neural Networkscited by 737
  10. 2017One pixel attack for fooling deep neural networkscited by 582
  11. 2019Graph U-Netscited by 492
  12. 2016Geometric deep learning on graphs and manifolds using mixture model CNNscited by 468
  13. 2020Parameterized Explainer for Graph Neural Networkcited by 426
  14. 2019Self-Attention Graph Poolingcited by 421
  15. 2021Beyond Low-frequency Information in Graph Convolutional Networkscited by 362
  16. 2020Learning to Learn Single Domain Generalizationcited by 296
  17. 2020The Risks of Invariant Risk Minimizationcited by 262
  18. 2020Hierarchical Generation of Molecular Graphs using Structural Motifscited by 208
  19. 2019Universal Invariant and Equivariant Graph Neural Networkscited by 183
  20. 2019Domain Generalization Using a Mixture of Multiple Latent Domainscited by 170
  21. 2019A Meta-Transfer Objective for Learning to Disentangle Causal Mechanismscited by 158
  22. 2021Deep Stable Learning for Out-Of-Distribution Generalizationcited by 153
  23. 2021Generative Causal Explanations for Graph Neural Networkscited by 124
  24. 2019Learning Robust Representations by Projecting Superficial Statistics Outcited by 123
  25. 2021Does Invariant Risk Minimization Capture Invariance?cited by 116
  26. 2015Reducing Overfitting in Deep Networks by Decorrelating Representationscited by 94
  27. 2020A PAC-Bayesian Approach to Generalization Bounds for Graph Neural Networkscited by 85
  28. 2016Discovering Causal Signals in Imagescited by 79
  29. 2019Stable Learning via Sample Reweightingcited by 70
  30. 2020Stable Prediction with Model Misspecification and Agnostic Distribution Shiftcited by 62
  31. 2019Test-Time Training with Self-Supervision for Generalization under Distribution Shiftscited by 59
  32. 2018Stable Prediction across Unknown Environmentscited by 55
  33. 2021Knowledge-Preserving Incremental Social Event Detection via Heterogeneous GNNscited by 51
  34. 2016Approximate Residual Balancing: De-Biased Inference of Average Treatment Effects in High Dimensionscited by 50
  35. 2016Regularizing CNNs with Locally Constrained Decorrelationscited by 41
  36. 2017Causally Regularized Learning with Agnostic Data Selection Biascited by 29
  37. 2019Learning Backtrackless Aligned-Spatial Graph Convolutional Networks for Graph Classificationcited by 13
  38. 2020BiTe-GCN: A New GCN Architecture via BidirectionalConvolution of Topology and Features on Text-Rich Networkscited by 7

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year

Citing papers per year: 2022: 4; 2023: 8; 2024: 10; 2025: 3.

25 papers.

Sort:

  1. 2022Out-Of-Distribution Generalization on Graphs: A Surveycited by 97
  2. 2022Trustworthy Graph Neural Networks: Aspects, Methods and Trendscited by 89
  3. 2022Debiasing Graph Neural Networks via Learning Disentangled Causal Substructurecited by 83
  4. 2024Deep Graph Anomaly Detection: A Survey and New Perspectivescited by 40
  5. 2024Alleviating Structural Distribution Shift in Graph Anomaly Detectioncited by 32
  6. 2023Data-centric Graph Learning: A Surveycited by 22
  7. 2023Causality and Independence Enhancement for Biased Node Classificationcited by 13
  8. 2023Predicting the Silent Majority on Graphs: Knowledge Transferable Graph Neural Networkcited by 10
  9. 2024Graph Fairness Learning under Distribution Shiftscited by 10
  10. 2024Learning Invariant Representations of Graph Neural Networks via Cluster Generalizationcited by 10
  11. 2023Graph Out-of-Distribution Generalization with Controllable Data Augmentationcited by 9
  12. 2024Graph Contrastive Invariant Learning from the Causal Perspectivecited by 6
13 more
  1. 2022Directed Acyclic Graph Structure Learning from Dynamic Graphscited by 5
  2. 2023When Graph Neural Network Meets Causality: Opportunities, Methodologies and An Outlookcited by 4
  3. 2024Topology-Aware Dynamic Reweighting for Distribution Shifts on Graphcited by 3
  4. 2024A Survey of Out-of-distribution Generalization for Graph Machine Learning from a Causal Viewcited by 3
  5. 2025Causal invariant geographic network representations with feature and structural distribution shiftscited by 3
  6. 2023Deep Stable Multi-Interest Learning for Out-of-distribution Sequential Recommendationcited by 2
  7. 2024DeCaf: A Causal Decoupling Framework for OOD Generalization on Node Classificationcited by 2
  8. 2025Mitigating Message Imbalance in Fraud Detection with Dual-View Graph Representation Learningcited by 2
  9. 2025Evolving Graph Learning for Out-of-Distribution Generalization in Non-stationary Environmentscited by 2
  10. 2023FRGNN: Mitigating the Impact of Distribution Shift on Graph Neural Networks via Test-Time Feature Reconstructioncited by 1
  11. 2023Towards Human-like Perception: Learning Structural Causal Model in Heterogeneous Graphcited by 1
  12. 2024IENE: Identifying and Extrapolating the Node Environment for Out-of-Distribution Generalization on Graphscited by 0
  13. 2024Contextual Representation Anchor Network to Alleviate Selection Bias in Few-Shot Drug Discoverycited by 0