Workshops Archive

Intelligent Sampling of Extreme-Scale Turbulence Datasets for Accurate and Efficient Spatiotemporal Model Training

Workshop: AI4S: 6th Workshop on Artificial Intelligence and Machine Learning for Scientific Applications

Authors: Wesley Brewer, Murali Meena Gopalakrishnan, Matthias Maiterth, Aditya Kashi, Jong Youl Choi, Pei Zhang, and Stephen Nichols (Oak Ridge National Laboratory (ORNL)); Riccardo Balin (Argonne National Laboratory (ANL)); Miles Couchman (York University); Stephen de Bruyn Kops (University of Massachusetts Amherst); P.K. Yeung, Daniel Dotson, and Rohini Uma-Vaideswaran (Georgia Institute of Technology); and Sarp Oral and Feiyi Wang (Oak Ridge National Laboratory (ORNL))

Abstract: With the end of Moore's law and Dennard scaling, efficient training increasingly requires rethinking data volume. Can we train better models with significantly less data via intelligent subsampling? To explore this, we develop SICKLE, a sparse intelligent curation framework for efficient learning, featuring a novel maximum entropy (MaxEnt) sampling approach, scalable training, and energy benchmarking. We compare MaxEnt with random and phase-space sampling on large direct numerical simulation (DNS) datasets of turbulence. Evaluating SICKLE at scale on Frontier, we show that subsampling as a preprocessing step can, in many cases, improve model accuracy and substantially lower energy consumption, with observed reductions of up to 38x.

Back to AI4S: 6th Workshop on Artificial Intelligence and Machine Learning for Scientific Applications Archive Listing Back to Full Workshop Archive Listing