Data-driven synchronization-avoiding algorithms in the explicit distributed structural analysis of soft tissue

TL;DR

This paper introduces a data-driven approach using neural networks to reduce communication in distributed explicit finite element analysis of soft tissue, enhancing computational efficiency.

Contribution

It presents a novel synchronization-avoiding algorithm trained on solver data, improving efficiency in distributed soft tissue analysis.

Findings

The neural network accurately predicts shared node displacements.

The method reduces communication overhead in distributed computations.

Numerical experiments confirm the stability of the approach.

Abstract

We propose a data-driven framework to increase the computational efficiency of the explicit finite element method in the structural analysis of soft tissue. An encoder-decoder long short-term memory deep neural network is trained based on the data produced by an explicit, distributed finite element solver. We leverage this network to predict synchronized displacements at shared nodes, minimizing the amount of communication between processors. We perform extensive numerical experiments to quantify the accuracy and stability of the proposed synchronization-avoiding algorithm.