Researchers have developed a gated multimodal model to predict energy performance scores for residential buildings, integrating tabular data, free text descriptions, and GIS spatial features. This approach aims to provide scalable assessments for decarbonizing buildings, which contribute significantly to UK and EU emissions. The model achieved strong predictive accuracy in a London case study and demonstrated that combining multiple data types enhances performance over single-modality approaches. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT Provides a scalable framework for property-level energy efficiency assessment and retrofit planning, supporting net-zero housing transitions.
RANK_REASON The cluster contains an academic paper detailing a new multimodal learning model for property energy performance prediction.