RelEx: A Model-Agnostic Relational Model Explainer

Zhang, Yue; DeFazio, David; Ramesh, A.

doi:10.48550/arxiv.2006.00305

Cited by 7 publications

(8 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After that, Zhang et al [34] proposed a model-agnostic relational model explainer called RelEx, which treats the underlying model as a black-box model and learns relational explanations. The RelEx constructs explanations using two steps-learning a local differentiable approximation of the black-box model and then learning an interpretable mask over the local approximation with the use of subgraphs.…”

Section: Interpretable Gnn Through Subgraphsmentioning

confidence: 99%

See 1 more Smart Citation

A Survey of Explainable Graph Neural Networks: Taxonomy and Evaluation Metrics

Li¹,

Zhou²,

Verma³

et al. 2022

Preprint

View full text Add to dashboard Cite

Graph neural networks (GNNs) have demonstrated a significant boost in prediction performance on the graph data. At the same time, the predictions made by these models are often hard to interpret. In that regard, many efforts have been made to explain the prediction mechanisms of these models from perspectives such as GNNExplainer, XGNN and PGExplainer. Although such works present systematic frameworks to interpret GNNs, a holistic review for explainable GNNs is unavailable. In this survey, we present a comprehensive review of explainability techniques developed for GNNs. We focus on explainable graph neural networks, categorize them based on the use of explainable methods. We further provide the common performance metrics for GNNs explanations and point out several future research directions.

show abstract

Section: Interpretable Gnn Through Subgraphsmentioning

confidence: 99%

“…Robustness means the explanations of interpretation methods resist attacks such as input corruption/perturbation, adversarial attack and model manipulation. A robust interpretation method can provide similar explanations despite the presence of such attacks [11,34].…”

Section: Evaluation Metrics For Explaniable Techniquesmentioning

confidence: 99%

A Survey of Explainable Graph Neural Networks: Taxonomy and Evaluation Metrics

Li¹,

Zhou²,

Verma³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Moreover, we conduct ablation studies and sensitivity analysis in Appendix G to better understand the model components and validate the effectiveness of the designed objective. (Luo et al, 2020;Ying et al, 2019;Yuan et al, 2020a;Yue Zhang, 2020;Michael Sejr Schlichtkrull, 2021) learns the masks on graph features. Typically, GNN-Explainer (Ying et al, 2019) applies the instance-wise masks on the messages carried by graph structures, and maximizes the mutual information between the masked graph and the prediction.…”

Section: Study Of Generators (Rq2)mentioning

confidence: 99%

Deconfounding to Explanation Evaluation in Graph Neural Networks

Wu¹,

Wang²,

Zhang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Explainability of graph neural networks (GNNs) aims to answer "Why the GNN made a certain prediction?", which is crucial to interpret the model prediction. The feature attribution framework distributes a GNN's prediction to its input features (e.g., edges), identifying an influential subgraph as the explanation. When evaluating the explanation (i.e., subgraph importance), a standard way is to audit the model prediction based on the subgraph solely. However, we argue that a distribution shift exists between the full graph and the subgraph, causing the out-ofdistribution problem. Furthermore, with an in-depth causal analysis, we find the OOD effect acts as the confounder, which brings spurious associations between the subgraph importance and model prediction, making the evaluation less reliable. In this work, we propose Deconfounded Subgraph Evaluation (DSE) which assesses the causal effect of an explanatory subgraph on the model prediction. While the distribution shift is generally intractable, we employ the front-door adjustment and introduce a surrogate variable of the subgraphs. Specifically, we devise a generative model to generate the plausible surrogates that conform to the data distribution, thus approaching the unbiased estimation of subgraph importance. Empirical results demonstrate the effectiveness of DSE in terms of explanation fidelity.

show abstract

“…As discussed in Section 1 not much effort has been devoted to explainability in graph classification. Explanation methods for node classification and link prediction have been proposed in [16,43,44], but this is a different problem from the graph classification task in which the aim is to classify the whole graph and not its nodes.…”

Section: Related Workmentioning

confidence: 99%

Counterfactual Graphs for Explainable Classification of Brain Networks

Abrate

Bonchi

2021

Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery &Amp; Data Mining

View full text Add to dashboard Cite

Training graph classifiers able to distinguish between healthy brains and dysfunctional ones, can help identifying substructures associated to specific cognitive phenotypes. However, the mere predictive power of the graph classifier is of limited interest to the neuroscientists, which have plenty of tools for the diagnosis of specific mental disorders. What matters is the interpretation of the model, as it can provide novel insights and new hypotheses.In this paper we propose counterfactual graphs as a way to produce local post-hoc explanations of any black-box graph classifier. Given a graph and a black-box, a counterfactual is a graph which, while having high structural similarity with the original graph, is classified by the black-box in a different class. We propose and empirically compare several strategies for counterfactual graph search. Our experiments against a white-box classifier with known optimal counterfactual, show that our methods, although heuristic, can produce counterfactuals very close to the optimal one. Finally, we show how to use counterfactual graphs to build global explanations correctly capturing the behaviour of different black-box classifiers and providing interesting insights for the neuroscientists.

show abstract

RelEx: A Model-Agnostic Relational Model Explainer

Cited by 7 publications

References 9 publications

A Survey of Explainable Graph Neural Networks: Taxonomy and Evaluation Metrics

A Survey of Explainable Graph Neural Networks: Taxonomy and Evaluation Metrics

Deconfounding to Explanation Evaluation in Graph Neural Networks

Counterfactual Graphs for Explainable Classification of Brain Networks

Contact Info

Product

Resources

About