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:20260731T133318Z
LOCATION:Bldg. 8 - B 102
DTSTART;TZID=Europe/Stockholm:20260629T133000
DTEND;TZID=Europe/Stockholm:20260629T153000
UID:submissions.pasc-conference.org_PASC26_sess111@linklings.com
SUMMARY:MS1H - From Physics to Data - and Back: Trustworthy Machine Learni
 ng Potentials for Materials Design
DESCRIPTION:Access the recording\n\nOrganizer(s): Michał Sanocki, Ian Stör
 mer, Julija Zavadlav Koller (Technical University of Munich), and Philip R
 obin Loche (EPFL)\n\nMachine learning interatomic potentials (MLIPs) are t
 ransforming electronic-structure modeling by delivering near–quantum mecha
 nical accuracy at greatly reduced cost. As these models increasingly surro
 gate explicit electronic-structure calculations, especially for large, com
 plex, or heterogeneous systems, the central challenge becomes establishing
  trust when direct quantum-mechanical validation is no longer feasible. Ba
 lancing data-driven flexibility with physical consistency is essential: sy
 mmetry-preserving and physics-informed architectures enhance reliability, 
 while more flexible models demand rigorous verification to ensure they lea
 rn the correct physical relationships. Exascale computing reshapes this la
 ndscape by enabling large-scale data generation, systematic cross-validati
 on, and interrogation of MLIP behavior under extreme conditions. At the sa
 me time, the complexity of training and deploying these models highlights 
 the need for reproducibility, explainability, and robust uncertainty quant
 ification. Emerging approaches aim to embed trust checks - such as enforci
 ng conservation laws, equivariance, and stability-directly into training r
 ather than relying solely on post hoc validation. This minisymposium will 
 gather researchers from chemistry, materials science, physics, and compute
 r science to discuss strategies for developing trusted MLIPs. Topics inclu
 de physics-informed representations, equivariant architectures, long-range
  and non-conservative interactions, uncertainty estimation, and automated 
 verification workflows. The session aims to chart pathways toward reliable
 , scalable, and transparent MLIP frameworks for molecular and materials si
 mulation.\n\nFrom Dataset-Induced Biases to Uncertainty-Driven Learning of
  Interatomic Potentials\n\nMachine-learned interatomic potentials have sig
 nificantly advanced atomistic modeling by combining near first-principles 
 accuracy with the efficiency of classical potentials. However, their relia
 bility in simulations strongly depends on the datasets used for training, 
 both for specialized and founda...\n\n\nViktor Zaverkin (Saarland Universi
 ty, Leibniz Institute for New Materials)\n---------------------\nQuantum (
 Bio) Molecular Simulations with Machine Learning Force Fields\n\nMachine l
 earning force fields (MLFFs) promise to bridge the gap between quantum-mec
 hanical accuracy and the computational efficiency needed to simulate reali
 stic (bio)molecular systems [1]. Yet their predictive power is often limit
 ed by the quality and coverage of training data, as well as by locali...\n
 \n\nAdil Kabylda (University of Luxembourg)\n---------------------\nEnd-to
 -End Uncertainty Quantification for Atomistic Machine Learning\n\nMachine-
 learning interatomic potentials (MLIPs) have found widespread adoption in 
 atomistic simulation workflows. Their use, however, introduces statistical
  uncertainties from finite training data and approximation errors, as well
  as systematic discrepancies inherited from the underlying electronic-...\
 n\n\nMatthias Kellner (EPFL)\n---------------------\nOpen Discussion: Trus
 tworthy Machine Learning Potentials - Challenges and Perspectives\n\nThis 
 session will be an open discussion following the presentations in the mini
 symposium. The goal is to create space for speakers and participants to ex
 change perspectives on common challenges in developing and applying machin
 e learning interatomic potentials. Topics raised by the talks include dat.
 ..\n\n\nMichał Sanocki (Technical University of Munich)\n\nDomain: Chemist
 ry and Materials, Physics, Computational Methods and Applied Mathematics\n
 \nSession Chairs: Michał Sanocki (Technical University of Munich); Ian Stö
 rmer (Technical University of Munich); and Philip Loche (Technical Univers
 ity of Munich, EPFL)
END:VEVENT
END:VCALENDAR