2026-10-13T09:57:44Z https://www.scidb.cn/oai

10.57760/sciencedb.0109x 2026-08-31T10:15:18Z

2026-08-31 A Neuro-Symbolic Guardrail for Compliance with Aircraft Noise Laws: Source Code and Experiment Archive LSP LS doi:10.57760/sciencedb.0109x en This dataset is the reproducibility archive accompanying the manuscript "Compliance with Aircraft Noise Laws across Jurisdictions: Automated Equivalence Checking with a Neuro-Symbolic Guardrail" (under review at IEEE Access). It contains: (1) the complete Python source code (regulation-as-code limit engine, type-design cross-reference, six-step adjudicator, LLM extraction pipeline); (2) all experiment artifacts (per-case adjudications, per-call JSONL logs, machine-readable summaries for seven experiments); (3) the original UK CAA Aircraft Noise Database (CAAdB) files; (4) ground-truth tables parsed from EASA TCDSN documents. Coverage: 6,664 Chapter 3 jet configurations (19,992 pointwise limit checks), 11,688/4,754 Chapter 4/14 cumulative records, 99 real TCDSN extraction cases, four-seed × 80-case fault-injection experiments, and a five-model comparison. aircraft noise; airworthiness certification; neuro-symbolic AI; runtime assurance; large language models; regulation-as-code; computational comparative regulation zheng ping xin EMBARGO https://creativecommons.org/licenses/by/4.0/ dataset Science Data Bank