BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260421T090513Z
LOCATION:Bldg. 6 - Room 003
DTSTART;TZID=Europe/Stockholm:20260629T133000
DTEND;TZID=Europe/Stockholm:20260629T153000
UID:submissions.pasc-conference.org_PASC26_sess103@linklings.com
SUMMARY:MS1B - Applications of AI and ML Towards Addressing Magnetic Fusio
 n Challenges
DESCRIPTION:Nuclear fusion is increasingly seen as a credible complement t
 o renewable energy, driven today by both public and growing private invest
 ments. The leading fusion concepts, tokamaks and stellarators, rely on mag
 netic confinement. Fusion plasmas exhibit multiscale electromagnetic insta
 bilities, from machine-scale disruptions to microscopic turbulence. The de
 velopment of these systems thus requires extensive high-performance comput
 ing to understand their complex physics and to address the engineering cha
 llenges. Fully kinetic models are prohibitively expensive, so hierarchies 
 of reduced models are needed. Machine learning has therefore become a powe
 rful tool, enabling for example fast surrogate models for plasma equilibri
 um solvers and transport modules. Such models are also developed for build
 ing sub-system modules integrated into comprehensive physics-engineering d
 igital twin environments. In some cases, optimization is such that these m
 odels can be integrated into real-time control systems. These different ap
 plications of machine learning to magnetic fusion challenges will be addre
 ssed in this minisymposium. The talks will provide an overview of the adva
 nced ML techniques applied and should thus be of interest to a broad audie
 nce.\n\nPedestal Inference Engine (PIE): ML facilitated simulation-based i
 nference framework for tokamak pedestals\n\nThe increase of readily availa
 ble computing resources and advancement of simulations-based inference (SB
 I) algorithms is opening a pathway to bring statistical inference to the f
 orefront in model validation, discovery, and prediction in science and tec
 hnology [1]. In the context of tokamak fusion pl...\n\n\nAaro Järvinen, Am
 anda Bruncrona, Daniel Jordan, Adam Kit, and Anna Niemelä (VTT); Laurent C
 hone (CSC - IT Center for Science); Lorenzo Frassinetti and Arnaud Lafay (
 KTH Royal Institute of Technology); Alex Panera-Alvarez, Sven Wiesen, Tawe
 esak Jitsuk, and MJ Pueschel (DIFFER); Vlado Menkovski and Kiet Bennema te
 n Brinke (Technical University of Eindhoven); Samuli Saarelm and Lorenzo Z
 anisi (UKAEA); Tobias Görler (IPP Garching); Michele Marin (EPFL); and Dav
 id Hatch (University of Texas at Austin)\n---------------------\nScaling a
 nd Scarcity: Developing Robust ML Surrogates for Next-Generation Fusion Wo
 rkflows\n\nSurrogate modelling is emerging as a key paradigm in fusion res
 earch. It provides fast, reliable software components to accelerate high-f
 idelity plasma simulators and digital twins. However, their development is
  intrinsically linked to data availability and parameterization. This work
  explores data-...\n\n\nSamy Kerboua-Benlarbi (EPFL); Christoph Angerer (N
 VIDIA Inc.); Florian Cabot and Francesco Carpanese (EPFL); Jonathan Citrin
  (DeepMind); Hammam Elaian (EPFL); Cyrille Favreau (NVIDIA Inc.); Federico
  Felici (DeepMind); Davide Fransos (NVIDIA Inc.); Philippe Hamel (DeepMind
 ); Holger Reimerdes, Cristian Sommariva, and Elena Tonello (EPFL); Felix Y
 ang (NVIDIA Inc.); and Alessandro Pau and Olivier Sauter (EPFL)\n---------
 ------------\nSurrogate models of nonlinear gyrokinetics and high-fidelity
  fluid-kinetic edge plasma codes\n\nHigh-fidelity gyrokinetic turbulence m
 odelling and scrape-off layer modelling are crucial for scenario developme
 nt, but they not included in integrated models due to the associated prohi
 bitive computational cost. Firstly, this talk will present surrogate model
 s of nonlinear gyrokinetics (operating i...\n\n\nLorenzo Zanisi (UKAEA)\n-
 --------------------\nThe PORTALS Workflow: Accelerating Core Transport Pr
 edictions with Surrogate-Based Optimization Methods\n\nPredicting core per
 formance in magnetic confinement fusion devices requires accurately modeli
 ng the balance between heating sources and transport mechanisms. However, 
 the nonlinear electromagnetic turbulence in tokamak and stellarator plasma
 s makes such modeling computationally demanding, as capturi...\n\n\nP. Rod
 riguez-Fernandez and N.T. Howard (MIT Plasma Science and Fusion Center) an
 d J. Candy (General Atomics)\n\nDomain: Physics, Computational Methods and
  Applied Mathematics\n\nSession Chairs: Stephan Brunner (EPFL), Eric Sonne
 ndruecker (Max Planck Institute for Plasma Physics), and Florian Hindenlan
 g (Max Planck Institute for Plasma Physics)
END:VEVENT
END:VCALENDAR