Workshop: AI-DRIVEN FACADE DESIGN LAB

From Facade Recognition to HyPar Optimization

Date, Duration & Location

Saturday, 6 December 2025 | 09:00 – 18:00 (Full-Day Workshop)

Yeditepe University, Lab 222

Organizers

Instructor: Instr. MSc. Architect Özden Başak Suvakçı, MSc. Data Scientist Mustafa Çelen

Department: Architecture, Yeditepe University

Workshop Description

This one-day workshop introduces architecture students to deep learning with PyTorch and its application in sustainable design.

Students will train a façade segmentation model using Berkeley’s Facade Dataset (Pix2Pix dataset from UC Berkeley), extract measurable architectural data such as window-to-wall ratio (WWR) and vertical rhythm, and use these insights to inform optimization of tile placement on a hyperbolic paraboloid (HyPar) surface.

The session connects data-driven reasoning with architectural form, emphasizing how AI-based analysis can support resource efficiency, reduce waste, and enhance design sustainability.

Application link: https://forms.gle/Z9ADp3QupJF8rpsS7