In this 20 minutes webinar, we will compare Deep-Learning based approaches to traditional reduced order modeling or response surfaces, today used in the industry. We first will look at the differences from a purely technical perspective with an emphasis on performance and capabilities. This analysis will help us understand the ongoing revolution towards the democratization of advanced analytics and optimization in engineering.

Like other learning-based methods Geometric Convolutional Neural Networks can be used to build surrogate models of numerical solvers. However, it suffers none of the drawbacks of previous surrogate methods. It is agnostic to the shape parameters as it processes directly the mesh representation of the design. Hence, optimization or design parameters are decoupled from the learning problem and a single predictor can be trained with a large amount of data and used for many optimization tasks.

Unlike Kriging methods, the engineer does not have to choose and stick to a specific parametrization from the beginning to the end of experiments. Optimizations can be naturally warm-started using data from previous campaigns. Furthermore, it can leverage on transfer learning abilities of Deep models to blend simulations from multiple sources and with multiple fidelities.