Note: This video may require joining the NVIDIA Developer Program or login

GTC Silicon Valley-2019 ID:S91014:Understanding Buildings Using Machine Learning

JeanPaul Baloche(NOVALIAN),Constantin Chevance(NOVALIAN),Javier Venegas(NOVALIAN)
Well talk about data exploration and how its enabling next-generation buildings. Well describe how were using a mesh-based server application to tackle the building industrys headaches, and how GPU acceleration with Python allows us to perform large calculations. Our data scraping and data wrangling processes include geometrical transformation functions, high-level matrix multiplication, proximity search algorithm in quadtree structures, and advanced topological functions. Well explain how we use GPUs to train a neural network to accurately classify information within a building to create a clean and exploitable data model, a task thats impossible using the actual systems. Well also cover how this application benefits architects, technical coordinators, project owners, and others by increasing control accuracy and optimizing time and costs.

View the slides (pdf)