Deep Learning-Based Feature-Aware Data Modeling for Complex Physics Simulations
Machine Learning
2019-12-10 v1 Computer Vision and Pattern Recognition
Image and Video Processing
Abstract
Data modeling and reduction for in situ is important. Feature-driven methods for in situ data analysis and reduction are a priority for future exascale machines as there are currently very few such methods. We investigate a deep-learning based workflow that targets in situ data processing using autoencoders. We propose a Residual Autoencoder integrated Residual in Residual Dense Block (RRDB) to obtain better performance. Our proposed framework compressed our test data into 66 KB from 2.1 MB per 3D volume timestep.
Cite
@article{arxiv.1912.03587,
title = {Deep Learning-Based Feature-Aware Data Modeling for Complex Physics Simulations},
author = {Qun Liu and Subhashis Hazarika and John M. Patchett and James Paul Ahrens and Ayan Biswas},
journal= {arXiv preprint arXiv:1912.03587},
year = {2019}
}
Comments
Accepted as a research poster at the International Conference for High Performance Computing, Networking, Storage, and Analysis (SC19)