New SeeExplainer method offers faithful GNN explanations by capturing synergistic edge effects

Researchers have introduced SeeExplainer, a novel parameter-free method designed to provide faithful explanations for graph neural networks (GNNs). Unlike previous approaches that often overlook the combined impact of multiple edges, SeeExplainer captures these synergistic effects by first refining the graph into disjoint granular-balls. These balls are then used to construct a structural graph, allowing for a more accurate assessment of edge importance through node and edge perturbations. Experiments on various graph classification datasets demonstrate that SeeExplainer surpasses existing state-of-the-art explanation methods. AI

IMPACT This new method could improve the interpretability of GNNs, aiding in debugging and building trust in AI systems that use graph data.

RANK_REASON The cluster contains a research paper detailing a new method for explaining graph neural networks. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• graph classification datasets
• graph neural networks
• SeeExplainer

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →