Bianca Grizhar

Using Bayesian inference on building performance data to assist early stage design.

Coming from a background in Artificial Intelligence, Bianca applies machine learning techniques to the challenge of building performance analysis in early stage building design.

This research focusses on Bayesian belief networks that can structure causality between fixed, uncertain and unknown parameters. Bianca’s goal is to build a system that assists building scientists and architectural designers in making early design decisions based on experience and knowledge gathered from existing buildings.

Supervisors

Mike Donn (Architecture/Building Science) and Marcus Frean (Engineering/Computer Science).