Pre-registration (frozen 9 Nov 2026): replication of the ACM over-smoothing advantage (arXiv 2407.19231) for GCN and SGC on Cora and Citeseer, as released and with every method's learning rate, weight decay and dropout tuned equally
Frozen protocol, decision code, job design, harness, planted-effect table and kernel for a private pre-registered replication, deposited before any confirmatory run. Clause: the mean relative accuracy improvement of ACM over the strongest of None, BatchNorm, PairNorm and DropEdge across ten backbone-dataset-depth cells (published 20.75%), in an as-released arm and an equal-tuning arm. Verdicts are Replicated / Did not replicate / Inconclusive under protocol X.
Authors
- John Goodman (ORCID: https://orcid.org/0009-0008-4080-8043)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-09
- DOI
- https://doi.org/10.5281/zenodo.24092853
- Primary Topic
- Advanced Graph Neural Networks
- Type
- article
- Field-Weighted Citation Impact
- 0.00