# Model-Level GNN Explanations via Rule-to-Graph Readout for Logit Reconstruction

arXiv: https://arxiv.org/abs/2503.09051 · HTML: https://mlsanity.com/paper/2503.09051/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 36 references that are in the map.
- First posted on arXiv 12 Mar 2025.

## Builds on (most cited first, 21 of 21)

- [Semi-Supervised Classification with Graph Convolutional Networks](https://mlsanity.com/paper/1609.02907/) (2016)
- [Inductive Representation Learning on Large Graphs](https://mlsanity.com/paper/1706.02216/) (2017)
- [How Powerful are Graph Neural Networks?](https://mlsanity.com/paper/1810.00826/) (2018)
- [Modeling Relational Data with Graph Convolutional Networks](https://mlsanity.com/paper/1703.06103/) (2017)
- [Gated Graph Sequence Neural Networks](https://mlsanity.com/paper/1511.05493/) (2015)
- [GNNExplainer: Generating Explanations for Graph Neural Networks](https://mlsanity.com/paper/1903.03894/) (2019)
- [Parameterized Explainer for Graph Neural Network](https://mlsanity.com/paper/2011.04573/) (2020)
- [On Explainability of Graph Neural Networks via Subgraph Explorations](https://mlsanity.com/paper/2102.05152/) (2021)
- [XGNN: Towards Model-Level Explanations of Graph Neural Networks](https://mlsanity.com/paper/2006.02587/) (2020)
- [Generative Causal Explanations for Graph Neural Networks](https://mlsanity.com/paper/2104.06643/) (2021)
- [Global Explainability of GNNs via Logic Combination of Learned Concepts](https://mlsanity.com/paper/2210.07147/) (2022)
- [GStarX: Explaining Graph Neural Networks with Structure-Aware Cooperative Games](https://mlsanity.com/paper/2201.12380/) (2022)
- [GNNInterpreter: A Probabilistic Generative Model-Level Explanation for Graph Neural Networks](https://mlsanity.com/paper/2209.07924/) (2022)
- [Global Concept-Based Interpretability for Graph Neural Networks via Neuron Analysis](https://mlsanity.com/paper/2208.10609/) (2022)
- [DEGREE: Decomposition Based Explanation For Graph Neural Networks](https://mlsanity.com/paper/2305.12895/) (2023)
- [1-WL Expressiveness Is (Almost) All You Need](https://mlsanity.com/paper/2202.10156/) (2022)
- [Explaining Graph Neural Networks via Structure-aware Interaction Index](https://mlsanity.com/paper/2405.14352/) (2024)
- [GOAt: Explaining Graph Neural Networks via Graph Output Attribution](https://mlsanity.com/paper/2401.14578/) (2024)
- [MAGE: Model-Level Graph Neural Networks Explanations via Motif-based Graph Generation](https://mlsanity.com/paper/2405.12519/) (2024)
- [PAGE: Prototype-Based Model-Level Explanations for Graph Neural Networks](https://mlsanity.com/paper/2210.17159/) (2022)
- [EiG-Search: Generating Edge-Induced Subgraphs for GNN Explanation in Linear Time](https://mlsanity.com/paper/2405.01762/) (2024)

## Abstract (arXiv)

We propose a novel model-level GNN explanation framework that shifts the explanation target from class-wise rule extraction to rule-based logit reconstruction. Our method recasts the graph-level readout of a pretrained GNN as a weighted rule-level readout: grounded subgraph concepts are composed into logical rules, rule embeddings are computed directly from their symbolic structure, and active rules are passed through the frozen classifier head to reconstruct the GNN's raw multiclass logits. As a result, our approach provides global explanations that remain instantiable on unseen graphs, support subgraph-level grounding, and admit rule-level contribution analysis at test-time. Experiments on three synthetic and two real-world graph classification benchmarks show that our approach faithfully reconstructs the base GNN's raw multiclass logits, achieving high probability-level fidelity across datasets. Rule-level ablations further demonstrate that the identified critical rules actively support the predicted class while suppressing non-target classes, suggesting that they act as functional units rather than merely serving as post-hoc symbolic artifacts. Compared with prior class-wise rule-based explainers, our approach achieves competitive or better prediction agreement while being up to (20×) faster, and additionally provides rule weights, test-time grounding, and logit-level contribution analysis.