- By:
- Brewer, Wesley H; Gopalakrishnan Meena, Muralikrishnan ; Zhang, Pei ; Kashi, Aditya ; Choi, Jong Youl ; Maiterth, Matthias ; Nichols, Dudley S; Balin, Riccardo; de Bruyn Kops, Stephen; Dotson, Daniel; Yeung, P.K.; Vaideswaran, Rohini; Wang, Feiyi
- Page Number:
- 1-10
- Book Title:
- Proceedings of the SC '25 Workshops of the International Conference for High Performance Computing, Networking, Storage and Analysis
- Publication Date:
- March 12, 2026
- Publisher Location:
- Association for Computing Machinery, New York, New York, United States of America
- Conference Name:
- The International Conference for High Performance Computing, Networking, Storage, and Analysis
- Conference Location:
- St. Louis, Missouri, United States of America
- Conference Sponsor:
- ACM
- View DOI Listing:
- https://doi.org/10.1145/3731599.3767340
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 38×.
