CONSONANCE.for your information
Monday, 5 October 2026frenvi

Worth reading closely

01 — agents 43 upvotes

Architect-Ant: Editable Automatic Furnishing of Architectural Floor Plans

QUESTION — Can professional furnishing knowledge be learned from real floor plans using a pretrained model, enabling direct constraint-aware layout generation without relying on costly iterative agentic inference?

This work introduces AntPlan, a curated dataset of 505 real professional architectural floor plans with dense furniture annotations spanning 92 object classes and ten residential room categories, and Architect-Ant, a framework for generating furniture layouts. Architect-Ant represents layouts with an editable coordinate-based DSL and first learns professional furnishing patterns through supervised fine-tuning. It is then optimized with GRPO using a Layout Rule Score (LRS) that aggregates geometric and functional constraints derived from professional plans, providing outcome-level supervision without prescribed reasoning traces. Experiments show that Architect-Ant combines low geometric violation rates with high functional completeness while remaining object-level editable and convertible into 3D scenes.

AntPlan is a curated dataset of 505 real professional architectural floor plans with dense furniture annotations spanning 92 object classes and ten residential room categories.

OldDelorean · 30 Sept 2026 read the original ↗
↑