Machine Learning for Architectural Floor Plan Evaluation

The research focuses on computational methods for assessing architectural floor plan quality, with particular emphasis on machine learning and multimodal models. It investigates different evaluation paradigms, including similarity to reference designs, compliance with architectural requirements, intrinsic design quality, and instruction-following. A central focus is text-floor-plan alignment- evaluating whether architectural layouts correspond to requirements expressed in natural language. These requirements range from explicitly computable properties, such as room dimensions, spatial adjacency, and connectivity, to perceptual qualities, such as architectural style, historical character, and spatial experience. The research combines systematic analysis of existing methods with the development of architecture-specific datasets, benchmarks, and automated evaluation approaches using vision–language models and computational verification. Its broader objective is to establish reliable methods for evaluating architectural design quality and supporting more controllable, requirement-driven generative AI systems.