PATTERN Can Be Solved from Memory: Generator-Aware Baselines for Synthetic Node Classification

Name: PATTERN Can Be Solved from Memory: Generator-Aware Baselines for Synthetic Node Classification
Uploaded: 2026-10-18T23:09:08.641565+00:00
Duration: 333 s
Description: The PATTERN benchmark tests how well models recognize hidden subgraphs in graphs. A simple matcher, using only training labels, achieves 99.2% accuracy without training, beating published models like CKGConv (89.8%). This suggests that pattern recognition can be achieved through clever inference on the generator's structure rather than deep model training.

The PATTERN benchmark tests how well models recognize hidden subgraphs in graphs. A simple matcher, using only training labels, achieves 99.2% accuracy without training, beating published models like CKGConv (89.8%). This suggests that pattern recognition can be achieved through clever inference on the generator's structure rather than deep model training.

Open the full episode