Deep Learning for Scientific Computing: Recent Success, challenges, and next steps

Brian Van Essen
Lawrence Livermore National Laboratory

In this talk we will present some of the recent work in the field and at LLNL to combine large scale deep learning with scientific simulations, notably a simulation on inertial confinement fusion. In addition to the challenges of training large models, we will present a simulation workflow that enables us to create sufficient training data set sizes, and how this work will come together in the future. We will also touch on several other success in applying large scale deep learning.

Back to Science at Extreme Scales: Where Big Data Meets Large-Scale Computing Tutorials