One connector, four scholarly indexes, real DOIs and legal free PDFs. Works in Claude and ChatGPT.
Paste it into Claude or ChatGPT as a custom connector. 3-day free trial.
Jonas Degrave, Federico Felici, Jonas Buchli, Michael Neunert, Brendan Tracey, F. Carpanese, Timo Ewalds, Roland Hafner, Abbas Abdolmaleki, Diego de Las Casas and 21 more
Nuclear fusion using magnetic confinement, in particular in the tokamak configuration, is a promising path towards sustainable energy. A core challenge is to shape and maintain a high-temperature plasma within the tokamak vessel. This requires high-dimensional, high-frequency, closed-loop control using magnetic actuator coils, further complicated by the diverse requirements across a wide range of plasma configurations. In this work, we introduce a previously undescribed architecture for tokamak magnetic controller design that autonomously learns to command the full set of control coils. This architecture meets control objectives specified at a high level, at the same time satisfying physical and operational constraints. This approach has unprecedented flexibility and generality in problem specification and yields a notable reduction in design effort to produce new plasma configurations. We successfully produce and control a diverse set of plasma configurations on the Tokamak à Configuration Variable 1,2 , including elongated, conventional shapes, as well as advanced configurations, such as negative triangularity and ‘snowflake’ configurations. Our approach achieves accurate tracking of the location, current and shape for these configurations. We also demonstrate sustained ‘droplets’ on TCV, in which two separate plasmas are maintained simultaneously within the vessel. This represents a notable advance for tokamak feedback control, showing the potential of reinforcement learning to accelerate research in the fusion domain, and is one of the most challenging real-world systems to which reinforcement learning has been applied.
Formatted from the source metadata, not typed by hand. For a whole reference list, use the citation generator or ask your assistant through Find Research.
Degrave, J., Felici, F., Buchli, J., Neunert, M., Tracey, B., Carpanese, F., Ewalds, T., Hafner, R., Abdolmaleki, A., Casas, D. D. L., Donner, C., Fritz, L., Galperti, C., Huber, A., Keeling, J., Tsimpoukelli, M., Kay, J., Merle, A., Moret, J., … Riedmiller, M. (2022). Magnetic control of tokamak plasmas through deep reinforcement learning. Nature, 602(7897), 414–419. https://doi.org/10.1038/s41586-021-04301-9
Degrave, Jonas, et al. "Magnetic control of tokamak plasmas through deep reinforcement learning." Nature, vol. 602, no. 7897, 2022, pp. 414–419, https://doi.org/10.1038/s41586-021-04301-9.
Degrave, Jonas, Federico Felici, Jonas Buchli, Michael Neunert, Brendan Tracey, F. Carpanese, Timo Ewalds, et al. "Magnetic control of tokamak plasmas through deep reinforcement learning." Nature 602, no. 7897 (2022): 414–419. https://doi.org/10.1038/s41586-021-04301-9.
Degrave J., Felici F., Buchli J. [и др.] Magnetic control of tokamak plasmas through deep reinforcement learning // Nature. — 2022. — Vol. 602, № 7897. — P. 414–419. — URL: https://doi.org/10.1038/s41586-021-04301-9
@article{degrave2022magnetic,
title = {{Magnetic control of tokamak plasmas through deep reinforcement learning}},
author = {Jonas Degrave and Federico Felici and Jonas Buchli and Michael Neunert and Brendan Tracey and F. Carpanese and Timo Ewalds and Roland Hafner and Abbas Abdolmaleki and Diego de Las Casas and Craig Donner and Leslie Fritz and C. Galperti and Andrea Huber and James Keeling and Maria Tsimpoukelli and Jackie Kay and A. Merle and J.M. Moret and Seb Noury and Federico Pesamosca and David Pfau and O. Sauter and Cristian Sommariva and S. Coda and B.P. Duval and A. Fasoli and Pushmeet Kohli and Koray Kavukcuoglu and Demis Hassabis and Martin Riedmiller},
journal = {Nature},
year = {2022},
volume = {602},
number = {7897},
pages = {414--419},
doi = {10.1038/s41586-021-04301-9},
url = {https://doi.org/10.1038/s41586-021-04301-9}
}TY - JOUR AU - Degrave, Jonas AU - Felici, Federico AU - Buchli, Jonas AU - Neunert, Michael AU - Tracey, Brendan AU - Carpanese, F. AU - Ewalds, Timo AU - Hafner, Roland AU - Abdolmaleki, Abbas AU - Casas, Diego de Las AU - Donner, Craig AU - Fritz, Leslie AU - Galperti, C. AU - Huber, Andrea AU - Keeling, James AU - Tsimpoukelli, Maria AU - Kay, Jackie AU - Merle, A. AU - Moret, J.M. AU - Noury, Seb AU - Pesamosca, Federico AU - Pfau, David AU - Sauter, O. AU - Sommariva, Cristian AU - Coda, S. AU - Duval, B.P. AU - Fasoli, A. AU - Kohli, Pushmeet AU - Kavukcuoglu, Koray AU - Hassabis, Demis AU - Riedmiller, Martin TI - Magnetic control of tokamak plasmas through deep reinforcement learning JO - Nature VL - 602 IS - 7897 SP - 414 EP - 419 PY - 2022 DA - 2022/02/16 DO - 10.1038/s41586-021-04301-9 UR - https://doi.org/10.1038/s41586-021-04301-9 ER -
Add Find Research to Claude or ChatGPT and ask in plain language. The assistant reads this record, finds the papers around it and cites only what it actually found:
Setup takes about a minute: add the connector, then start the free trial.
Record from OpenAlex; open-access location from Unpaywall and the indexes. Find Research links only to legal copies.