mlsanityData to 14 Oct 2026

Sound Logical Explanations for Mean Aggregation Graph Neural Networks

arXiv 2511.11593 ↗First posted 27 Oct 2025cs.LG cs.AI cs.LOcited by 2

Not on the map. Why?

Abstract (arXiv)

Graph neural networks (GNNs) are frequently used for knowledge graph completion. Their black-box nature has motivated work that uses sound logical rules to explain predictions and characterise their expressivity. However, despite the prevalence of GNNs that use mean as an aggregation function, explainability and expressivity results are lacking for them. We consider GNNs with mean aggregation and non-negative weights (MAGNNs), proving the precise class of monotonic rules that can be sound for them, as well as providing a restricted fragment of first-order logic to explain any MAGNN prediction. Our experiments show that restricting mean-aggregation GNNs to have non-negative weights yields comparable or improved performance on standard inductive benchmarks, that sound rules are obtained in practice, that insightful explanations can be generated in practice, and that the sound rules can expose issues in the trained models.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 29 of its 50 references that are in the map. First posted on arXiv 27 Oct 2025.

Builds on

29 of 50 references.

References · 29

  1. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  2. 2017Graph Attention Networkscited by 8,923
  3. 2017Inductive Representation Learning on Large Graphscited by 6,320
  4. 2017Neural Message Passing for Quantum Chemistrycited by 3,487
  5. 2019A Comprehensive Survey on Graph Neural Networkscited by 2,919
  6. 2017Modeling Relational Data with Graph Convolutional Networkscited by 1,694
  7. 2019RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Spacecited by 898
  8. 2020Knowledge Graphscited by 621
  9. 2019Inductive Relation Prediction by Subgraph Reasoningcited by 239
  10. 2019DRUM: End-To-End Differentiable Rule Mining On Knowledge Graphscited by 175
  11. 2018RippleNet: Propagating User Preferences on the Knowledge Graph for Recommender Systemscited by 138
  12. 2020RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphscited by 118
17 more
  1. 2020BoxE: A Box Embedding Model for Knowledge Base Completioncited by 105
  2. 2022Global Explainability of GNNs via Logic Combination of Learned Conceptscited by 65
  3. 2019Leveraging Knowledge Bases in LSTMs for Improving Machine Readingcited by 37
  4. 2023The Descriptive Complexity of Graph Neural Networkscited by 26
  5. 2023Some Might Say All You Need Is Sumcited by 21
  6. 2024Decidability of Graph Neural Networks via Logical Characterizationscited by 14
  7. 2018A Recurrent Graph Neural Network for Multi-Relational Datacited by 13
  8. 2024Logical Characterizations of Recurrent Graph Neural Networks with Reals and Floatscited by 13
  9. 2023On the Correspondence Between Monotonic Max-Sum GNNs and Datalogcited by 12
  10. 2024A Logic for Reasoning About Aggregate-Combine Graph Neural Networkscited by 10
  11. 2024Covered Forest: Fine-grained generalization analysis of graph neural networkscited by 10
  12. 2024Almost Surely Asymptotically Constant Graph Neural Networkscited by 7
  13. 2024Logical Distillation of Graph Neural Networkscited by 7
  14. 2025Logical Characterizations of GNNs with Mean Aggregationcited by 6
  15. 2024Utilizing Description Logics for Global Explanations of Heterogeneous Graph Neural Networkscited by 4
  16. 2023Revisiting Inferential Benchmarks for Knowledge Graph Completioncited by 2
  17. 2024Relational Graph Convolutional Networks Do Not Learn Sound Rulescited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Structural Preservation and the Logical Expressiveness of Graph Neural Networkscited by 0
  2. 2026Beyond Trained Models: Compiling GNNs for a Sound Explainer Benchmarkcited by 0