Explaining GNN-based Recommendations in Logic
Summary: Introduce Makex, a logic-based explanation framework that discovers Rules for ExPlanations (REPs)—a graph pattern Q plus precondition predicates X→M(x,y)—to expose topology and feature dependencies driving GNN recommender outputs. Defines REPs via 1-WL, gives algorithms for global REP discovery and top-k local explanations, and reports superior fidelity, sparsity, and efficiency versus prior methods. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Wenfei Fan (Beihang University; Shenzhen University; University of Edinburgh)
- 2. Lihang Fan (Beihang University)
- 3. Dandan Lin (Shenzhen University)
- 4. Min Xie (Shenzhen University)
BibTeX Citation
@article{fan_vldb25,
title = {{Explaining GNN-based Recommendations in Logic}},
author = {Fan, Wenfei and Fan, Lihang and Lin, Dandan and Xie, Min},
journal = {PVLDB},
series = {{VLDB} '25},
volume = {18},
number = {3},
pages = {715--728},
doi = {10.14778/3712221.3712237},
url = {https://doi.org/10.14778/3712221.3712237},
year = {2025}
}Incoming Citations (Sorted by Pagerank)
Showing 1 of 1 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 10,716 | Efficient GNN Training on Giant Graphs with Collective Batching and Scheduling | 2026 | VLDB | 4.9769913e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 10 of 10 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,369 | word2vec, node2vec, graph2vec, X2vec: Towards a Theory of Vector Embeddings of Structured Data | 2020 | PODS | 0.00010899682 |
| 1,943 | Interpretable Data-Based Explanations for Fairness Debugging | 2022 | SIGMOD | 9.3254131e-05 |
| 2,365 | Dependencies for Graphs | 2017 | PODS | 8.5625389e-05 |
| 3,027 | Functional Dependencies for Graphs | 2016 | SIGMOD | 7.7351194e-05 |
| 4,860 | Association Rules with Graph Patterns | 2015 | VLDB | 6.3763445e-05 |
| 6,484 | On Data-Aware Global Explainability of Graph Neural Networks | 2023 | VLDB | 5.767533e-05 |
| 8,225 | Towards Event Prediction in Temporal Graphs | 2022 | VLDB | 5.3737187e-05 |
| 8,299 | Capturing Associations in Graphs | 2020 | VLDB | 5.3587491e-05 |
| 9,813 | Making It Tractable to Catch Duplicates and Conflicts in Graphs | 2023 | SIGMOD | 5.1233734e-05 |
| 11,730 | Enriching Recommendation Models with Logic Conditions | 2023 | SIGMOD | 4.9769913e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 10,986 | GraphAgent: An Effective Knowledge-Guided GNN Model Selection System | 2026 | VLDB |
| 2 | 10,749 | Database Views as Explanations for Relational Deep Learning | 2026 | VLDB |
| 3 | 3,098 | FlexRecs: Expressing and Combining Flexible Recommendations | 2009 | SIGMOD |
| 4 | 4,844 | REX: Explaining Relationships between Entity Pairs | 2012 | VLDB |
| 5 | 11,769 | HENCE-X: Toward Heterogeneity-agnostic Multi-level Explainability for Deep Graph Networks | 2023 | VLDB |
| 6 | 4,549 | Explaining Link Prediction Systems based on Knowledge Graph Embeddings | 2022 | SIGMOD |
| 7 | 11,339 | Lighter-X: An Efficient and Plug-and-play Strategy for Graph-based Recommendation through Decoupled Propagation | 2025 | VLDB |
| 8 | 6,484 | On Data-Aware Global Explainability of Graph Neural Networks | 2023 | VLDB |
| 9 | 11,730 | Enriching Recommendation Models with Logic Conditions | 2023 | SIGMOD |
| 10 | 10,109 | View-based Explanations for Graph Neural Networks | 2024 | SIGMOD |