TokaMind: A Multi-Modal Transformer Foundation Model for Tokamak Plasma Dynamics DOI: XXXXXXX.XXXXXXX Conference: 32nd SIGKDD Conference on Knowledge Discovery and Data Mining, 2026; August 09–13, 2026; Jeju, Korea ISBN: XXX-X-XXXX-XXXX-X/2026/08

Tobia Boschi, Andrea Loreti, Nicola C. Amorisco, Rodrigo H. Ordonez-Hurtado, Cécile Rousseau, George K. Holt, Eszter Székely, Alexander Whittle, Samuel Jackson, Adriano Agnello, Stanislas Pamela, Alessandra Pascale, Robert Akers, Juan Bernabe Moreno, Vassil Alexandrov, Mykhaylo Zayats · 2026

Read TokaMind: A Multi-Modal Transformer Foundation Model for Tokamak Plasma Dynamics DOI: XXXXXXX.XXXXXXX Conference: 32nd SIGKDD Conference on Knowledge Discovery and Data Mining, 2026; August 09–13, 2026; Jeju, Korea ISBN: XXX-X-XXXX-XXXX-X/2026/08 by Tobia Boschi, Andrea Loreti, Nicola C. Amorisco, Rodrigo H. Ordonez-Hurtado, Cécile Rousseau, George K. Holt, Eszter Székely, Alexander Whittle, Samuel Jackson, Adriano Agnello, Stanislas Pamela, Alessandra Pascale, Robert Akers, Juan Bernabe Moreno, Vassil Alexandrov, Mykhaylo Zayats (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-15084.txt.

Description

Boschi and colleagues present TokaMind, billed as the first open-source foundation model for tokamak plasma dynamics. It rests on a multi-modal transformer pretrained on mixed diagnostics from the public MAST data set. The model handles time series, 2D profiles, and videos at different sampling rates, copes with missing signals, and adapts to new tasks by loading and freezing four model parts in selective fashion.

Signals are encoded with a light fixed-basis discrete cosine transform embedding, with a clean interface for alternatives such as variational autoencoders. On the TokaMark benchmark of 14 reconstruction and forecasting tasks, fine-tuned TokaMind beats the strongest baseline on all but one task. Warm-start adaptation helps most on hard downstream settings such as long-horizon forecasting and high-dimensional equilibrium goals, and code plus weights are public.

Catalog record

CREATORTobia Boschi, Andrea Loreti, Nicola C. Amorisco, Rodrigo H. Ordonez-Hurtado, Cécile Rousseau, George K. Holt, Eszter Székely, Alexander Whittle, Samuel Jackson, Adriano Agnello, Stanislas Pamela, Alessandra Pascale, Robert Akers, Juan Bernabe Moreno, Vassil Alexandrov, Mykhaylo Zayats
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-15084

Browse the full library · About the Tower of Babel