Geometric Deep Learning for Product Design // Applied AI Meetup August 2024

This event is sponsored by Recursive Awesome (AI Strategy) & Lab651 (Custom AI Software Development)
This month, we'll be diving into using Geometric Deep Learning in Product Design with Fatma Koçer-Poyraz! Physics-based simulations have been used successfully to speed up product design. These simulations predict the NVH, durability, aerodynamics, and electromagnetics performance of products among others. However, these simulations require lengthy model preparation and lengthy computations on expensive hardware. Hence design exploration that is required in concept design phases is not as accessible, leaving engineers with few design choices with trial-and-error to pursue rather than finding innovative, optimal designs that can be found through design exploration.
One of the long-asked questions have been whether historical simulation data can be used to train machine learning models to be used for fast design exploration. However, this data is most likely not parametric, neither parametrizable. This is the challenge geometric deep learning solves. With geometric deep learning, we no longer need parametric or parametrizable datasets to train ML models. In this presentation, applications of geometric deep learning will be shown along with how it is productized in Altair physicsAI.
Our speaker this month is Fatma Koçer-Poyraz, who has received a B.Sc. degree in Civil Engineering from the Middle East Technical University, Ankara, Turkey, and MSc and PhD degrees from the University of Iowa in Structural Optimization. After graduation, she joined Visteon to develop processes for design optimization of structural and electromagnetic parts. In 2006, she started in Altair as Product Line Manager for Altair HyperWorks design exploration and optimization products. Currently she works as the Vice President for Engineering Data Science. In this role, she and her team work on engineering data science strategy, development, and execution which includes investigating and applying the latest technologies in the field, providing feedback to Altair software, and supporting customer projects. Fatma Kocer-Poyraz is one of the recipients of Crain’s 2019 Notable Women in STEM recognition. She is also the Vice-Chair for NAFEMS Engineering Data Science Working Group since 2021.