Researchers have developed IfcLLM, a novel framework designed to make Industry Foundation Classes (IFC) data more accessible through natural language queries. The system converts IFC models into both relational and graph representations, which are then processed by an LLM with iterative reasoning to understand user requests. Implemented using an open-weight GPT OSS 120B model, the framework demonstrated high accuracy in tests, suggesting a path toward more intuitive interaction with complex BIM data. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT This framework could significantly improve accessibility for non-expert users in the AEC industry by enabling natural language interaction with complex BIM data.
RANK_REASON The cluster describes a research paper introducing a new framework and its evaluation.