Artificial intelligence tools bring efficiency to the architectural process
22 de outubro de 2024By Karina Toledo, from Bologna | FAPESP Innovative R&D – The conventional work of an architect is usually slow and artisanal. When hired to design an interior for a company, it can take days or even weeks, depending on the size of the project, to produce a few options to offer to the client. However, artificial intelligence tools can make this process much faster and more efficient.
This is what Arqgen, one of the startups supported by FAPESP’s Innovative Research in Small Businesses Program (PIPE), has done. “We’ve created a series of algorithms that solve architectural problems using artificial intelligence and generative design techniques. Our goal is to return architecture to its leading role, which is very strategic for any type of business,” Alexandre Kuroda, one of the company’s founders, explained to Agência FAPESP.
To date, Arqgen has two main products. One is aimed precisely at solving this issue of interior design, and its main clients are Itaú and Bradesco, the two largest private banks in Brazil. “These banks have architectural manuals with a series of rules for maintaining the standard of two branches. We ‘algorithmized’ these rules and combined them with artificial intelligence and generative design techniques, so that it became possible to produce all the possible layout options for a given property in less than half a day,” said the entrepreneur.
The second solution developed by the startup uses the same algorithms but applied to real estate development. In this case, the tool is used to determine the best layout for a building – whether residential or commercial – within a parcel of land.
The main results of the work were presented by Kuroda on October 15th in the city of Bologna. His company was one of the four PIPEs selected to take part in FAPESP Week Italy. The event, which began on October 14th, was organized in partnership with the Alma Mater Studiorum - Università di Bologna (UNIBO).
Alexandre Kuroda of Arqgen, a startup that translates corporate architecture manuals into algorithms and combines them with artificial intelligence and generative design techniques (photo: Karina Toledo/Agência FAPESP)
Another entrepreneur who took part in the symposium was Renata Bonaldi, founder of SleepUp. With the support of FAPESP (projects 22/13130-9 and 20/00666-2), she has developed several products to help diagnose and treat sleep disorders, mainly insomnia and obstructive sleep apnea. These include a headband that monitors sleep cycles at home with the same accuracy as a hospital polysomnography test and an app that uses principles of cognitive behavioral therapy (CBT) to help clients improve their sleep hygiene.
Renata Bonaldi of SleepUp, a company that develops products to help diagnose and treat sleep disorders (photo: Karina Toledo/Agência FAPESP)
Also taking part were Diogo Gará Caetano from Kasko and Rodrigo Basile Junqueira from Ocellott.
Kasco R&D is a spin-off company of the State University of Campinas (UNICAMP), focused on developing innovative systems for digital communication and solutions based on artificial intelligence and machine learning for various purposes, such as monitoring power lines, quality control in industry and recognizing beef cattle in images captured by drones.
Ocellott, which was also present at FAPESP Week China, focuses on aeronautical electrification (read more at: agencia.fapesp.br/52083).
The four entrepreneurs had the opportunity to present their businesses during a panel discussion on technology and innovation, which discussed various aspects related to Industry 5.0 – the industry that seeks synergy between human work and personalization with the efficiency and automation of Industry 4.0.
The panel included Mario Sergio Salerno, from the University of São Paulo (USP), and Luca Foschini and Franco Callegati, both from UNIBO. The event was coordinated by FAPESP’s General Coordinator Technologies and Innovation Partnerships, Rodolfo Azevedo, from UNICAMP, and Antonio Corradi, from UNIBO.