Graph Compression for Interpretable Graph Neural Network Inference At Scale
Summary: ExGIS performs a one-time compression G→G_c that preserves exact outputs of any L‑layer GNN on any node, enabling inference directly on the compressed graph without decompression. It parallelizes queries, returns concise high‑fidelity explanatory subgraphs, and offers interactive/LLM-driven natural-language explanations for scalable interpretable GNN inference. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Yangxin Fan
- 2. Haolai Che
- 3. Mingjian Lu
- 4. Yinghui Wu
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 2 of 2 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 2,177 | Accelerating Large Scale Real-Time GNN Inference using Channel Pruning | 2021 | VLDB | 9.359876e-05 |
| 9,764 | View-based Explanations for Graph Neural Networks | 2024 | SIGMOD | 4.2856106e-05 |
Previous
Page 1 / 1
Next