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:20260605T154541Z
LOCATION:Bldg. 8 - Room B 102
DTSTART;TZID=Europe/Stockholm:20260701T090000
DTEND;TZID=Europe/Stockholm:20260701T110000
UID:submissions.pasc-conference.org_PASC26_sess157@linklings.com
SUMMARY:MS4H - Agentic Workflows for Trustworthy Discovery in Materials Sc
 ience and Chemistry
DESCRIPTION:Organizer(s): Jan Janssen (Max Planck Institute for Sustainabl
 e Materials)\n\nAdvances in agentic AI—autonomous, tool-using systems that
  plan, call simulators, reason over uncertainty, and adapt—are poised to t
 ransform how we explore vast chemical and materials design spaces. This mi
 nisymposium will showcase state-of-the-art methods that couple agentic dec
 ision-making with scalable HPC workflows to accelerate hypothesis generati
 on, simulation throughput, and autonomy while strengthening scientific tru
 st. Topics span agent-driven hypothesis generation and assessment, automat
 ic execution of large computational campaigns based on different simulatio
 n scales, and workflow/runtime systems that autonomously schedule thousand
 s to millions of tasks across heterogeneous supercomputers with robust pro
 venance and reproducibility. A central thread is Building Trust in Science
  through HPC Co-Design: contributors will detail how agents, numerical met
 hods, software stacks, data services, are co-designed to deliver validated
 , reproducible, and/or auditable results using autonomous loops.\n\nSelf-V
 alidating Research Assistant for High Performance Electronic Structure Cal
 culations\n\nAgentic AI systems based on large language models (LLMs) offe
 r promising avenues for automating atomistic modeling workflows on high-pe
 rformance computing (HPC) platforms. However, incorrect workflow specifica
 tion in this context can lead to substantial computational waste and unrel
 iable scientific ...\n\n\nLuigi Genovese (CEA Grenoble)\n-----------------
 ----\npyiron – A Workflow Framework for Trustworthy Agentic Workflows in M
 aterials Science\n\nThe hierarchical nature of materials requires simulati
 on approaches that couple methods across disciplines and length scales. In
 tegrating heterogeneous simulation codes poses significant interoperabilit
 y challenges due to incompatible units, file formats, and data structures.
  The pyiron workflow fra...\n\n\nJan Janssen (Max-Planck-Insitute for Sust
 ainable Materials)\n---------------------\nAccelerating Molecular Discover
 y with AI Agents\n\nChemistry underpins many fields critical to modern soc
 iety, yet the rational design of chemical systems with targeted properties
  remains a formidable challenge due to the immense size of chemical space.
   Typical examples include the development of selective reagents for metal
  separations and robust ...\n\n\nDanny Perez (Los Alamos National Laborato
 ry)\n---------------------\nSemantic Provenance for Trustworthy Agentic Wo
 rkflows in Materials Science\n\nAgentic AI systems, autonomous agents capa
 ble of planning simulations, invoking computational tools, and reasoning o
 ver results, offer new opportunities for accelerating discovery in materia
 ls science. However, ensuring reproducibility of research endeavors remain
 s a key challenge when autonomous sy...\n\n\nEdan Bainglass, Xing Wang, Al
 exander Goscinski, Julian Geiger, and Giovanni Pizzi (Paul Scherrer Instit
 ute)\n\nDomain: Chemistry and Materials, Climate, Weather, and Earth Scien
 ces, Engineering, Physics\n\nSession Chair: Jan Janssen (Max Planck Instit
 ute for Sustainable Materials)
END:VEVENT
END:VCALENDAR