From ionic columns to twisting skyscrapers, architecture has changed drastically over the centuries. However, people can still recognize the beauty of the Parthenon just as easily as that of an elegant home in the Chicago suburbs. It can be difficult to quantify why a certain family is drawn to a certain house or how home designers appeal to certain homebuyers. The Artificial Intelligence (AI), Architecture Aesthetics and House Price Creative Inquiry project led by Dr. Lily Shen in the Department of Finance uses machine learning to determine how the beauty of a home affects its resale value.

Although beauty is not a simple thing to quantify, Shen and her team are employing machine learning to establish an automated method for determining the value of a property. The team looks at four elements in the machine learning algorithms: the overall balance and symmetry of the house; the landscape of the exterior; the type of roofline; and the proportionality of the windows, regarding the amount and combinations of styles.

“In the housing market, potential buyers don’t just look at the number of bedrooms, baths and square footage when searching for a new home, they also look at the unique qualities that the house has,” Megan Quinan, a senior accounting major, said.

For each of the four elements, students created a scale from one to five to determine how attractive the home might be to the average person. For example, relating to proportionality of the windows, if there were multiple window styles present on the house, this would result in a lower rating. Students provide points of reference for the computer to learn from such as the height of the ceiling or the size of a window. With each trial, the machine becomes better at predicting the initial assumptions. “It’s not magic! It really is efficient learning,” Shen said.

This project has captured the attention of finance and engineering students as well as policy makers working in the banking sector who are keen to explore how technology can help investors monitor financial risks. With housing market fluctuations, this research has the power to inform investors and bankers about the increased or decreased values of homes. In the future, this Creative Inquiry project hopes to become more financially-focused by answering questions about investment securities and institutional buyers.

Understandably, a machine cannot account for every aspect of someone’s attraction to a certain house, so Shen highlights the importance of human and AI interactions within this project. “[This] research is more about how humans or investors, companies or individuals, can use the latest technology to create value and change the world,” Shen said.

With the help of AI, technology can provide insight about a wide variety of topics in a more efficient and standardized way. This Creative Inquiry team is demonstrating how machine learning can simulate the human perspective and quantify the abstract value of beauty within architecture.