Introduction
This bi-monthly seminar series explores real-world applications of physics-informed machine learning (Φ-ML) methods to the engineering practice. They cover a wide range of topics, offering a cross-sectional view of the state of the art on Φ-ML research, worldwide.
Participants have the opportunity to hear from leading researchers and learn about the latest developments in this emerging field. These seminars also offer the chance to identify and spark collaboration opportunities.
About the event
For decades, researchers have been studying efficient numerical methods to solve differential equations, most of them optimised for one-core processors. However, the trend in today's computing industry to design high-performance processors looking at parallel architectures. Parallelisation across time appears to be a promising way to provide more parallelism. In this talk, we'll introduce one of the main algorithms, the Parareal algorithm. We'll discuss how time-parallel methods can accelerate the solution of chaotic problems and examine how the chaotic nature affects convergence and speedup. We'll also present the moving-window (MoWi) algorithm, an original method for the efficient sampling of initial conditions of nonlinear chaotic systems. MoWi uses time-parallel integration as a subroutine, harnessing its properties to explore large swathes of the solution space, from which to recover statistical properties of the system, all while achieving parallel speedup. We'll outline the conditions under which this method can outperform traditional implementations of time-parallel algorithms. This presentation is intended to highlight some core aspects of MoWi and to engage with researchers who may contribute to developing a physics-informed ML approach for an adaptive version of MoWi.
