Following the 7.1 magnitude Puebla-Morelos earthquake on September 19th, 2017, SOM sent a team of engineers to Mexico City to contribute to post-disaster recovery efforts. The team documented building damage and provided technical support to local structural reconnaissance efforts.
During their analysis of the research gathered from this trip, the team turned to machine learning as a way to enhance post-earthquake reconnaissance capabilities. Machine learning allowed the team to identify and classify building damage based on photo documentation, rather than traditional manual assessments, thus making the process faster and more efficient.
In this video, Structural Engineering Partner Mark Sarkisian and structural engineer Samantha Walker discuss the development of SOM’s machine learning initiative and proprietary machine learning tools, from its origins following the earthquake to the ways in which it has been utilized since.
From construction reading and verification to building maintenance and design, machine learning reflects new opportunities and benefits for the AEC industry.
This video was filmed at SOM’s “Arte + Ingeniería + Arquitectura” exhibition at San Ildefonso in Mexico City. SOM’s machine learning initiative was launched with advisory consultation from Anthony Sarkis, who went on to found Diffgram, a machine learning platform.