Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 1 h ago
PreprintCC BY 4.0arXiv

Towards Large-Scale Heterogeneous Data Organization for Scientific Foundation Models: A Nuclear Fusion Case Study

  • Nathaniel Chen
  • Kouroche Bouchiat
  • Peter Steiner
  • Azarakhsh Jalalvand
  • SangKyeun Kim
  • Egemen Kolemen
Open at sourcePDFDOICite

Abstract

Training effective foundation models requires massive and organized datasets, yet scientific domains such as nuclear fusion present unique challenges due to largely heterogeneous and sparse data. Here we characterize the data used in developing such a model: with over 20 sensor types spanning 5 orders of magnitude in sampling rate, mixed tensor structures (point measurements, spectrograms, images), and nonstationary physics. We analyze our input complexity and discuss trade-offs between temporal context and frequency resolution. Our analysis provides a template for representing multi-modal fluctuation data at scale, with implications for both multi-modal control systems and nuclear fusion.

Related

  1. A foundation model for energy and radiation systems built on heterogeneous scientific interfaces
    Samrendra Roy, Tapas Tripura, Yoon Pyo Lee +2
    arXivPreprintarXiv licenceFull textComputational physicsarXiv
  2. Learning Nonlocal Interactions in Large-Scale Heterogeneous Metamaterials through Scientific Machine Learning
    Dex Doksoo Lee, Jehyeon Shin, Yue Yu +3
    arXivPreprintarXiv licenceFull textOptics & photonicsarXiv
  3. Human-in-the-Loop Meta Bayesian Optimization for Fusion Energy and Scientific Applications
    Ricardo Luna Gutierrez, Sahand Ghorbanpour, Rahman Ejaz +5
    arXivPreprintarXiv licenceFull textPlasma physicsarXiv
  4. Towards Stellarator Geometry Optimisation for Nuclear Fusion
    Tobias Weißberg, Moritz Heep, Zorah Lähner +1
    arXivPreprintCC BY 4.0Full textPlasma physicsarXiv
  5. Panda Diplomacy: Foundation Model Pre-training across Particle Imaging Detectors for High Energy and Nuclear Physics
    Samuel Young, César Jesús-Valls, Kazuhiro Terao
    arXivPreprintCC BY 4.0Full textHigh-energy & particle physicsarXiv

Cite

BibTeXRIS
BibTeX
@misc{chen2026towards,
  title = {{Towards Large-Scale Heterogeneous Data Organization for Scientific Foundation Models: A Nuclear Fusion Case Study}},
  author = {Nathaniel Chen and Kouroche Bouchiat and Peter Steiner and Azarakhsh Jalalvand and SangKyeun Kim and Egemen Kolemen},
  howpublished = {arXiv},
  year = {2026},
  month = aug,
  doi = {10.48550/arxiv.2608.27578},
  eprint = {2608.27578},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2608.27578},
  note = {CC BY 4.0},
}
RIS
TY  - UNPB
TI  - Towards Large-Scale Heterogeneous Data Organization for Scientific Foundation Models: A Nuclear Fusion Case Study
AU  - Nathaniel Chen
AU  - Kouroche Bouchiat
AU  - Peter Steiner
AU  - Azarakhsh Jalalvand
AU  - SangKyeun Kim
AU  - Egemen Kolemen
PY  - 2026
DA  - 2026/08/31
PB  - arXiv
DO  - 10.48550/arxiv.2608.27578
UR  - https://arxiv.org/abs/2608.27578
AB  - Training effective foundation models requires massive and organized datasets, yet scientific domains such as nuclear fusion present unique challenges due to largely heterogeneous and sparse data. Here we characterize the data used in developing such a model: with over 20 sensor types spanning 5 orders of magnitude in sampling rate, mixed tensor structures (point measurements, spectrograms, images), and nonstationary physics. We analyze our input complexity and discuss trade-offs between temporal context and frequency resolution. Our analysis provides a template for representing multi-modal fluctuation data at scale, with implications for both multi-modal control systems and nuclear fusion.
ER  -