Human-Machine Collaboration for the Visual Exploration and Analysis of High-Dimensional Spatial Simulation Ensembles

Mai Dahshan, Nicholas Polys, Leanna House, Karim Youssef, Ryan Pollyea

2024

Abstract

Continuous improvements in supercomputing have given scientists from various fields the ability to conduct large-scale multi-dimensional numerical simulation ensembles. A simulation ensemble involves running multiple simulations, each with slight variations in model settings, such as input parameters, initial conditions, or boundary values. Exploring and analyzing these ensembles facilitates understanding parameter sensitivity and the correlations between different ensemble members. To capture these relationships, visual analytical tools are used to extract important features from the ensemble. In many cases, however, these visualizations highlight the differences in the ensemble using aggregated or descriptive statistics, ignoring the correlations and local differences between different spatial regions, which could hinder the exploration process. This paper proposes a visual analytical approach, SpatialGLEE, to interactively explore the spatial variability in the simulation ensemble. The proposed approach uses Gaussian Process Regression (GPR) and Semantic Interaction (SI) to help scientists explore the impact of input parameters on the ensemble and find the commonalities and differences across ensemble members and regions of interest (ROI). GPR models the spatial correlation structure in the ensemble. The modeled data is then inputted into the visualization pipeline for analysis and exploration with SI. The effectiveness of SpatialGLEE is demonstrated using a real-life case study.

Paper Citation

in Harvard Style

Dahshan M., Polys N., House L., Youssef K. and Pollyea R. (2024). Human-Machine Collaboration for the Visual Exploration and Analysis of High-Dimensional Spatial Simulation Ensembles. In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 1: IVAPP; ISBN 978-989-758-679-8, SciTePress, pages 678-689. DOI: 10.5220/0012405100003660

in Bibtex Style

@conference{ivapp24,
author={Mai Dahshan and Nicholas Polys and Leanna House and Karim Youssef and Ryan Pollyea},
title={Human-Machine Collaboration for the Visual Exploration and Analysis of High-Dimensional Spatial Simulation Ensembles},
booktitle={Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 1: IVAPP},
year={2024},
pages={678-689},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012405100003660},
isbn={978-989-758-679-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 1: IVAPP
TI - Human-Machine Collaboration for the Visual Exploration and Analysis of High-Dimensional Spatial Simulation Ensembles
SN - 978-989-758-679-8
AU - Dahshan M.
AU - Polys N.
AU - House L.
AU - Youssef K.
AU - Pollyea R.
PY - 2024
SP - 678
EP - 689
DO - 10.5220/0012405100003660
PB - SciTePress