XGNN: Towards Model-Level Explanations of Graph Neural Networks

By Haonan Yuan, Jiliang Tang, Xia Hu, Shuiwang Ji. Knowledge Discovery and Data Mining, 2020

498 citations recorded by the source corpus.

Graphs neural networks (GNNs) learn node features by aggregating and combining neighbor information, which have achieved promising performance on many graph tasks. However, GNNs are mostly treated as black-boxes and lack human intelligible explanations. Thus, they cannot be fully trusted and used in certain application domains if GNN models cannot be explained. In this work, we propose a novel approach, known as XGNN, to interpret GNNs at the model-level. Our approach can provide high-level insights and generic understanding of how GNNs work. In particular, we propose to explain GNNs by training a graph generator so that the generated graph patterns maximize a certain prediction of the model. We formulate the graph generation as a reinforcement learning task, where for each step, the graph generator predicts how to add an edge into the current graph. The graph generator is trained via a policy gradient method based on information from the trained GNNs. In addition, we incorporate several graph rules to encourage the generated graphs to be valid. Experimental results on both synthetic and real-world datasets show that our proposed methods help understand and verify the trained GNNs. Furthermore, our experimental results indicate that the generated graphs can provide guidance on how to improve the trained GNNs.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/75c8466a0c1c3b9fe595efc83671984ef95bd679

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Parameterized Explainer for Graph Neural Network (2020, unknown): “XGNN provides model-level explanations without preserving the local fidelity [55].”
Explainability in Graph Neural Networks: A Taxonomic Survey (2020, methodology): “approaches are proposed to explain the predictions of GNNs [44], such as XGNN [45], GNNExplainer [46], The authors are with the Department of Computer Science and Engineering, Texas A&MUniversity, College Station, TX 77843USA.”
On Explainability of Graph Neural Networks via Subgraph Explorations (2021, methodology): “The XGNN can obtain graph patterns to explain GNNs but its explanations are not input-dependent and less precise.”
Higher-Order Explanations of Graph Neural Networks via Relevant Walks (2020, methodology): “Further recently proposed methods that map the GNN prediction to graph substructures include XGNN [26], Trap2 [27] and GraphMask [28].”
Graph Neural Networks in Recommender Systems: A Survey (2020, background): “Themethods can be divided into two categories: the instance-levelmethods provide example-specific explanations by identifying important input features for its prediction [3, 123, 194]; the model-level methods provide high-level interpretations and a generic understanding of how deep graph models wor”
Explainable AI Methods - A Brief Overview (2020, methodology): “Idea: The XGNN method [88] is a post-hoc method that operates on the model level, meaning that it does not strive to provide individual example-level explanations.”
Interpretable and Generalizable Graph Learning via Stochastic Attention Mechanism (2022, methodology): “Graph-SST2 (Socher et al., 2013; Yuan et al., 2020b) is a sentiment analysis dataset, where each text sequence in SST2 is converted to a graph.”
Self-Supervised Learning of Graph Neural Networks: A Unified Review (2021, background): “[31], [32], [33], [34], [35] and other related fields [36], [37] have beenwell studied andmade substantial progress.”
Towards multi-modal causability with Graph Neural Networks enabling information fusion for explainable AI (2021, methodology): “Another new approach for explanations has been introduced recently by Yuan et al. (2020) [65], called XGNN, for interpreting deep graph models at model level.”