ATLASky-AI_4D_Spatiotemporal_Aerospace_Benchmark

 This dataset integrates three distinct aerospace data sources—AddQual turbine-blade inspection reports, CAD-based geometric features, and NASA ASRS aviation incidents—into a unified 4D (X, Y, Z, Time) spatiotemporal format. Using GPT‑4O, a finetuned aerospace model, each domain’s raw text or 3d model was converted into structured JSON, preserving key measurements, relationships, and timestamps. The AddQual portion reflects anonymized but authentically verified dimensional checks by aerospace experts, while CAD and NASA data are publicly sourced and synthetically enhanced for multi-modal consistency. Together, these data streams offer a robust testbed for knowledge graph construction, multi-modal data fusion, and real-time verification pipelines—particularly the ATLASky-AI framework’s Ranked Multi-Modal Verification (RMMVe) and Adaptive Intelligence Cycle (AAIC). By capturing realistic anomalies, partial coordinates, and masked proprietary details, this dataset showcases how next-generation aerospace inspection systems can unify heterogeneous data into a cohesive, verifiable knowledge environment.o1