PREDICTIVE AI: SEMANTIC VS FEATURE SEGMENTATION: 2022-02 Pablo Lorenzo-Eiroa's Research on Artificial Intelligence activates differences between predictive AI and generative AI through semantic and visual feature segmentation training repositories implementing their own Big Data surveys. The research aims at redefining history through a new way of looking at space and the object of architecture. A new theory of architecture through Big Data emerges in which the new lenses of obserbation derive into a new theory of the object. Insights into the forensic survey, meassurement, segmentation and prediction enabled the research to find new boundaries and relationships between the work of Borromini's San Carlo and Rainaldi's SM in Campitelli. The prediction AI estimation is based on global repositories and personal research repositories designed to train machine vision feature recognition. The research identifies architectural features in the historic critical integration of Rainaldi of the work of Borromini and Palladio.

The project proposes to displace assumed heritage and deconstruct colonizations implicit in the history of architecture through new theories on the object. The project looks both into eurocentric heritage to deconstruct possible conflicts and also non eurocentric architecture in Latin America and parts of the global south.

EXANDING POSSIBLE PASTS TO REDEFINE ARCHITECTURE POSSIBLE FUTURES: The project develops a new theory on the object through analytical Big Data patterns and machine vision not possible for the naked eye, looking for non-intuitive Big Data patterns insights able to deconstruct conventional cultural readings of the project. By implementing AI the project aims to both expand the history of archtiecture through observations on the object that could not happen otherwise. Usually the history of architecture is defined by istorians that developed theories based on readings, historic documentation and other means such as measurement and observation. We propose to develop new insights redefining the existing theories of the object by forensic evidence, developing non intuitive expanded observations. We therefore propose a new theory of architecture recomipling and crissing references with all existing histories and expanding into new theories of architecture never drawn. 

AUGMENTED REALITY AND "DIGITAL TWINS": the project is both a research project and a project in itself rethinking the future from expanding the past. The project is part of e-Architects and Pablo Lorenzo-Eiroa's research to expand Big Data repositories for historic preservation including environmental concerns. The project aims at refunctionalizing existing heritage, buildings and cities in relation to augmented reality making them fully navigable online and to train AI repositories and ANN. Pablo Lorenzo-Eiroa's research also includes generative AI over surveyed point cloud repositories, developing unique artistic environments and projective creative immersive environments.

CREDITS: survey SM Campitelli by Carlo Rainaldi by Andrew Saunders with the assistance of Pablo Lorenzo-Eiroa and Mario Graziano Valenti (2015). The result point cloud of LiDAR laser scanning using FARO M70 scanner is processed by Pablo Lorenzo-Eiroa through multiple platforms and interfaces, including Scene and others (2015-2022). Some processing was done by Pablo Lorenzo-Eiroa PI with Rsearch Assistant Salma Kattass (2021-22). Point cloud 250M points; estimation prediction based on 50,000 points and training on a reduced repository of 1000 points through several cycles.

The research was published in the book of May of 2023.


AI machine vision semantic segmentation prediction

For more information and description of the research and findings refer to the book "Digital Signifiers in an Architecture of Information: From Big Data and Simulation to Artificial Intelligence", Pablo Lorenzo-Eiroa, Routledge, London 2023: