Dissertation Talk: Gabe Fierro: Self-Adapting Software for Cyberphysical Systems; Friday, April 30, 3 PM PST
April 30, 2021
Date: Friday, April 30, 2021
Over the course of my PhD, I have been building (with collaborators at UC Berkeley, UC San Diego, Carnegie Mellon, LBNL and NREL) a number of open-source tools and systems which are enabling practical and scalable data-driven research in smart buildings and cities. These tools and systems — which synthesize techniques from database systems, data integration and ontology design — include Brick, a rich metadata ontology for buildings which allows software to discover and contextualize relevant data; HodDB, an efficient graph query processor permitting the use of graph metadata in existing analytics pipelines; and Mortar, an open testbed and analytics platform which uses semantic metadata to enable reproducible data science on a public dataset of over 100 real buildings. I will describe how these tools and systems promote a paradigm of “self-adapting software” which reduces or eliminates the cost of deploying data-driven sustainable practices at scale, and illustrate current successes in smart buildings and electric grid contexts.