EDU-AI: a twofold machine learning model to support classroom layout generation

Purpose This study aims to present a twofold machine learning (ML) model, namely, EDU-AI, and its implementation in educational buildings. The specific focus is on classroom layout design, which investigated regarding of ML the early phases design. Design/methodology/approach introduces framework adopts generative adversarial networks (GAN) architecture Pix2Pix method. processes data collection, set preparation, training, validation evaluation for proposed model are presented. trained over two coupled sets layouts extracted from typical school project database Ministry National Education Republic Turkey validated with foreign boundaries. generated objectively evaluated through structural similarity method (SSIM). Findings EDU-AI generates despite use small set. Objective evaluations show that can provide satisfactory outputs given boundaries regardless shape complexity (reserved newly synthesized). Originality/value specifically contributes automation generation using ML-based algorithms. EDU-AI’s two-step enables zoning any boundary furnishing previously zone. also be used design phase projects other countries. It adapted architectural typologies involving footprint, relations.