Named Lineages

Graph-conditioned generative models for spatial layout and floorplan generation. Includes the influential Graph2Plan → House-GAN → House-GAN++ → Building-GNN chain plus the GNN Co-pilot and Alymani-Jabi BGR retrieval thread.

Knowledge graph + topological BIM applied to building energy and digital twins. Combines BOT, Akroyd Universal Digital Twin, and Topologic-stack energy-prediction work.

Topologic + BOT (Building Topology Ontology) and the chain of papers that build on those non-manifold-topology + ontology foundations.

BIM element-classification papers using IFC graphs feeding GCN/GAT for element-type prediction. The Collins et al. IFC encoding framework is the central enabling work.

Bridging BIM and city-scale GIS into unified knowledge graphs via BOT, IFC, and RDF/OWL frameworks. Enables multi-scale urban analytics.

The Zhao Tsinghua group's GNN-driven structural design line — shear wall layout, beam layout, frame structure design — derived from foundational GML primitives.

Construction safety ontologies + knowledge graphs supporting hazard identification, regulatory compliance, and emergency response.

Recent LLM-augmented knowledge-graph work — BIM-GPT hybrids, GraphRAG, ontology-aware language models for AEC.