Leakage-Safe DGNN Benchmark on Scientific Collaboration Networks (data & code)

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Reproducible benchmark and quantified leakage decomposition for dynamic graph neural networks on co-authorship networks. Core-20 paired disciplines + extended fields; pipeline, per-snapshot results, honest leaderboard, leakcheck audit tool, CausalGuard certification.