Reproduce “Declarative Explanations for Graph Neural Networks: A Demonstration” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Cibo Yu, Haitong Tang, Yinghui Wu, Arijit Khan, Tingting Zhu, Tingyang Chen, Xiangyu Ke
SliceGXQ provides a declarative interface for layer-wise GNN explanations. Users select a graph, model, target, and constraints; its backend slices the model and returns verified explanations. In a YelpChi example, node 239 changes from “Benign” to “Fraud” by layer 3; fine-tuning that layer raises test accuracy from 0.7416 to 0.8389, an illustrative result that does not establish causation or generality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.