'What Makes AI Research Replicable? Executable Knowledge Graphs as Scientific Knowledge Representations' arXiv 2510.17795 — methodology paper addresses AI research replication crisis through executable knowledge graphs
The arXiv 2510.17795 paper addresses the AI research replication challenge through executable knowledge graphs as scientific knowledge representations. Replication failure rates in AI research are documented at substantial levels; executable knowledge graphs provide methodology for representing AI research with replication-supporting structure built in.
The substantive piece is the executable-knowledge-graph methodology for replication infrastructure. Pre-paper AI research replication relied on text-paper descriptions plus open-source code releases plus model-weight artifacts. The combination produced substantial replication-failure rates because text descriptions are interpretation-dependent, code releases are environment-fragile, model weights have provenance ambiguity. Executable knowledge graphs provide structured replication-supporting representation that addresses each limitation.
The competitive read against the broader H2 2026 AI research infrastructure landscape is that replication-infrastructure methodology compounds with evaluation-infrastructure (Holistic Agent Leaderboard, Efficient Benchmarking) and field-baseline-characterization (multiple survey papers) into substantively better-organized research infrastructure than H1 2026 baseline supported. The H2 2026 to 2027 AI research community is investing systematically in infrastructure improvements that compound across all research direction outputs.
arXiv — What Makes AI Research Replicable? (2510.17795) → · arXiv — Holistic Agent Leaderboard →