mlsanityData to 14 Oct 2026

Towards Understanding Link Predictor Generalizability Under Distribution Shifts

arXiv 2406.08788 ↗First posted 13 Jun 2024cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

State-of-the-art link prediction (LP) models demonstrate impressive benchmark results. However, popular benchmark datasets often assume that training, validation, and testing samples are representative of the overall dataset distribution. In real-world situations, this assumption is often incorrect; uncontrolled factors lead new dataset samples to come from a different distribution than training samples. Additionally, the majority of recent work with graph dataset shift focuses on node- and graph-level tasks, largely ignoring link-level tasks. To bridge this gap, we introduce a novel splitting strategy, known as LPShift, which utilizes structural properties to induce a controlled distribution shift. We verify LPShift's effect through empirical evaluation of SOTA LP models on 16 LPShift variants of original dataset splits, with results indicating drastic changes to model performance. Additional experiments demonstrate graph structure has a strong influence on the success of current generalization methods. Source Code Available Here: https://github.com/revolins/LPShift

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 44 of its 69 references that are in the map. First posted on arXiv 13 Jun 2024.

Builds on

44 of 69 references.

References · 44

  1. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  2. 2015Domain-Adversarial Training of Neural Networkscited by 3,044
  3. 2019Invariant Risk Minimizationcited by 1,854
  4. 2016Variational Graph Auto-Encoderscited by 1,382
  5. 2020WILDS: A Benchmark of in-the-Wild Distribution Shiftscited by 1,375
  6. 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,247
  7. 2016Deep CORAL: Correlation Alignment for Deep Domain Adaptationcited by 1,200
  8. 2020In Search of Lost Domain Generalizationcited by 1,054
  9. 2018Link Prediction Based on Graph Neural Networkscited by 832
  10. 2020Out-of-Distribution Generalization via Risk Extrapolation (REx)cited by 742
  11. 2019Graph Neural Networks for Social Recommendationcited by 714
  12. 2019DropEdge: Towards Deep Graph Convolutional Networks on Node Classificationcited by 680
32 more
  1. 2021Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Predictioncited by 275
  2. 2022Improving Out-of-Distribution Robustness via Selective Augmentationcited by 202
  3. 2021A Fine-Grained Analysis on Distribution Shiftcited by 171
  4. 2022Handling Distribution Shifts on Graphs: An Invariance Perspectivecited by 164
  5. 2022Learning on Large-scale Text-attributed Graphs via Variational Inferencecited by 155
  6. 2020Labeling Trick: A Theory of Using Graph Neural Networks for Multi-Node Representation Learningcited by 149
  7. 2022GOOD: A Graph Out-of-Distribution Benchmarkcited by 141
  8. 2022Out-Of-Distribution Generalization on Graphs: A Surveycited by 97
  9. 2020Rethinking Importance Weighting for Deep Learning under Distribution Shiftcited by 93
  10. 2020Large-scale Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classificationcited by 86
  11. 2021Size-Invariant Graph Representations for Graph Classification Extrapolationscited by 84
  12. 2022Graph Neural Networks for Link Prediction with Subgraph Sketchingcited by 81
  13. 2022Neo-GNNs: Neighborhood Overlap-aware Graph Neural Networks for Link Predictioncited by 78
  14. 2023Evaluating Graph Neural Networks for Link Prediction: Current Pitfalls and New Benchmarkingcited by 69
  15. 2022Empowering Graph Representation Learning with Test-Time Graph Transformationcited by 67
  16. 2023Neural Common Neighbor with Completion for Link Predictioncited by 64
  17. 2022DrugOOD: Out-of-Distribution (OOD) Dataset Curator and Benchmark for AI-aided Drug Discovery -- A Focus on Affinity Prediction Problems with Noise Annotationscited by 54
  18. 2023Model Agnostic Sample Reweighting for Out-of-Distribution Learningcited by 44
  19. 2021Relating Graph Neural Networks to Structural Causal Modelscited by 41
  20. 2022OOD Link Prediction Generalization Capabilities of Message-Passing GNNs in Larger Test Graphscited by 38
  21. 2022Unleashing the Power of Graph Data Augmentation on Covariate Distribution Shiftcited by 38
  22. 2021Pooling Architecture Search for Graph Classificationcited by 36
  23. 2024Alleviating Structural Distribution Shift in Graph Anomaly Detectioncited by 32
  24. 2024Graph Out-of-Distribution Generalization via Causal Interventioncited by 31
  25. 2023Revisiting Link Prediction: A Data Perspectivecited by 29
  26. 2023LPFormer: An Adaptive Graph Transformer for Link Predictioncited by 17
  27. 2021Edge Proposal Sets for Link Predictioncited by 16
  28. 2023Pitfalls in Link Prediction with Graph Neural Networks: Understanding the Impact of Target-link Inclusion & Better Practicescited by 13
  29. 2022FakeEdge: Alleviate Dataset Shift in Link Predictioncited by 12
  30. 2023A Topological Perspective on Demystifying GNN-Based Link Prediction Performancecited by 11
  31. 2023Label Deconvolution for Node Representation Learning on Large-scale Attributed Graphs against Learning Biascited by 7
  32. 2021The Paradox of Second-Order Homophily in Networkscited by 3

Built on by

Not cited yet.