ECHO-Synth

ECHO synthetic propagation task — node-level regression on synthetic long-range graphs. The primary catalog metric is MAE.

ECHO synthetic propagation task — node-level regression on synthetic long-range graphs. The cataloged artifact reports 10,080 graphs.

Each synthetic or physical graph supplies local node/edge state and a target created by controlled long-range propagation. The exact signal channels and propagation parameters are part of the named ECHO task and should not be mixed across variants.

The catalog records these protocols or variants: Standard split. Evaluation uses MAE (lower is better). Exact masks or folds must come from the cited release.

Performance is sensitive to propagation radius, graph size, and train/test distance regime. Randomly mixing configurations can turn a long-range extrapolation task into interpolation; report the named variant and fixed generation parameters.