Simulation-Based Parallel Training
Preprint
Abstract
Numerical simulations are ubiquitous in science and engineering. Machine learning for science investigates how artificial neural architectures can learn from these simulations to speed up scientific discovery and engineering processes. Most of these architectures are trained in a supervised manner. They require tremendous amounts of data from simulations that are slow to generate and memory greedy. In this article, we present our ongoing work to design a training framework that alleviates those bottlenecks. It generates data in parallel with the training process. Such simultaneity induces a bias in the data available during the training. We present a strategy to mitigate this bias with a memory buffer. We test our framework on the multi-parametric Lorenz's attractor. We show the benefit of our framework compared to offline training and the success of our data bias mitigation strategy to capture the complex chaotic dynamics of the system.Keywords
All Related Versions
This publication has 12 references indexed in Scilit:
- Simple computational strategies for more effective physics-informed neural networks modeling of turbulent natural convectionJournal of Computational Physics, 2022
- Machine learning–accelerated computational fluid dynamicsProceedings of the National Academy of Sciences, 2021
- Data-Driven Learning for the Mori--Zwanzig Formalism: A Generalization of the Koopman Learning FrameworkSIAM Journal on Applied Dynamical Systems, 2021
- Data-driven predictions of the Lorenz systemPhysica D: Nonlinear Phenomena, 2020
- Shallow neural networks for fluid flow reconstruction with limited sensorsProceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2020
- Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equationsJournal of Computational Physics, 2019
- Deep learning for universal linear embeddings of nonlinear dynamicsNature Communications, 2018
- Learning in Nonstationary Environments: A SurveyIEEE Computational Intelligence Magazine, 2015
- GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputersSoftwareX, 2015
- DIRECT NUMERICAL SIMULATION: A Tool in Turbulence ResearchAnnual Review of Fluid Mechanics, 1998