Addressing computational challenges in physical system simulations with machine learning

TL;DR

This paper introduces a machine learning framework that reduces the need for costly physical simulations by generating accurate, simulation-like data, enabling more efficient exploration of physical systems such as earthquake physics and material development.

Contribution

The paper presents a novel two-step machine learning approach combining supervised prediction and reinforcement learning to generate simulation-like data, addressing computational challenges in physical system modeling.

Findings

Effective data generation for earthquake physics case study

Improved data efficiency in material development simulations

Reduced computational costs in physical system analysis

Abstract

In this paper, we present a machine learning-based data generator framework tailored to aid researchers who utilize simulations to examine various physical systems or processes. High computational costs and the resulting limited data often pose significant challenges to gaining insights into these systems or processes. Our approach involves a two-step process: initially, we train a supervised predictive model using a limited simulated dataset to predict simulation outcomes. Subsequently, a reinforcement learning agent is trained to generate accurate, simulation-like data by leveraging the supervised model. With this framework, researchers can generate more accurate data and know the outcomes without running high computational simulations, which enables them to explore the parameter space more efficiently and gain deeper insights into physical systems or processes. We demonstrate the effectiveness of the proposed framework by applying it to two case studies, one focusing on earthquake rupture physics and the other on new material development.