Position: Opportunities Exist for Machine Learning in Magnetic Fusion Energy
Authors: Lucas Spangher, Allen M. Wang, Andrew Maris, Myles Stapelberg, Viraj Mehta, Alex Saperstein, Stephen Lane-Walsh, Akshata Kishore Moharir, Alessandro Pau, Cristina Rea
ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | In this position paper, we highlight six key research challenges in the field of fusion energy that we believe should be research priorities for the Machine Learning (ML) community because they are especially ripe for ML applications: (1) disruption prediction, (2) simulation and dynamics modeling (3) resolving partially observed data, (4) improving controls, (5) guiding experiments with optimal design, and (6) enhancing materials discovery. For each problem, we give background, review past ML work, suggest features of future models, and list challenges and idiosyncrasies facing ML development. Ours is the first to attempt a comprehensive overview of ML in tokamak fusion and to address the ML community directly. |
| Researcher Affiliation | Academia | 1Plasma Science and Fusion Center, Massachusetts Institute of Technology, 77 Massachusetts Ave, Cambridge, MA 02139, USA 2Department of Electrical Engineering and Computer Science, Carnegie Mellon University, 5000 Forbes Avenue Pittsburgh, PA 15213, USA 3Robert H. Smith School of Business, University of Maryland, College Park, 7901 Regents Drive College Park, MD 20742-5025, USA 4Swiss Plasma Center, Swiss Federal Institute of Technology, Chemin du Barrage 16, 1015 Ecublens, Lausanne, Switzerland. |
| Pseudocode | No | The paper discusses various ML approaches and models but does not provide any pseudocode or algorithm blocks. |
| Open Source Code | No | This paper is a position paper reviewing existing work; it does not describe a new methodology for which source code would be provided. There is no statement about releasing code or links to a code repository by the authors. |
| Open Datasets | Yes | Open/FAIR data grant (National Science Foundation, 2022; Almada et al., 2020) and the DOE s Pu RE grant (D.O.E, 2021) are examples of better practices; for examples in fusion see (Montes et al., 2020). [...] Current datasets (Chanussot et al., 2021) for MLIPs are often insufficient for fusion-relevant materials [...] Datacomp: In search of the next generation of multimodal datasets. Ar Xiv, abs/2304.14108, 2023. |
| Dataset Splits | No | The paper is a position paper that reviews concepts and challenges; it does not conduct its own experiments or provide specific details on training, validation, or test dataset splits. |
| Hardware Specification | No | The paper refers to computational requirements and mentions hardware (e.g., 'millions of CPU-hours', 'GPU parallelization', 'V100 GPUs', 'Mac Book Pro') in the context of the reviewed literature or general discussion, not as the specific hardware used for any experiments conducted by the authors in this paper. |
| Software Dependencies | No | The paper discusses software concepts and tools relevant to fusion research (e.g., MDSplus, IMAS, MuJoCo, Isaac Gym/Sim) but does not list specific software dependencies with version numbers for any experimental setup or methodology presented by the authors. |
| Experiment Setup | No | The paper is a position paper that outlines research opportunities and challenges; it does not describe an experimental setup, including hyperparameters or system-level training settings, for any experiments conducted by the authors. |