BEGIN:VCALENDAR VERSION:2.0 PRODID:-//CERN//INDICO//EN BEGIN:VEVENT SUMMARY:Automated Data Generation for Machine Learning Force Fields with a ims-PAX and FHI-aims DTSTART:20261007T120000Z DTEND:20261008T153000Z DTSTAMP:20260910T024200Z UID:indico-event-23@events.ms1p.org CONTACT:aims-coordinators@ms1p.org DESCRIPTION:Speakers: Tobias Henkes (University of Luxembourg\, Luxembourg )\n\n\n\n \n\nSave the date(Calendar Entry / Zoom Link)\n\n\nWatch on You Tube\n\nWatch on KouShare\n\nJoin us for this webinar\, consisting of an e xpert talk and interactive hands-on sessions\, on automated data generatio n for machine learning force fields with aims-PAX and FHI-aims.\nMachine l earning force fields (MLFFs) have transformed molecular and materials simu lations. They have become the method of choice for simulating the dynamics of large and complex systems\, from flexible biomolecules to bulk materia ls\, with near quantum-chemical accuracy at a fraction of the computationa l cost of ab initio methods. Yet the accuracy and reliability of any MLFF depends upon the data it is trained on. Collecting representative\, high-q uality reference data sets remains a bottleneck. It is often labor-intensi ve\, requires expert knowledge\, and consumes substantial computational re sources.\nActive learning has proven to be a successful strategy for addre ssing this data challenge\, allowing the model itself to decide which conf igurations are most informative and should be labeled with expensive ab in itio calculations. In this webinar\, we will explore how aims-PAX (Paralle l Active eXploration)\, a fully automated\, open-source framework integrat ed with FHI-aims\, streamlines the data generation process. Tobias Henkes will showcase how aims-PAX autonomously builds high-quality reference data sets by coupling diversified multi-trajectory sampling with scalable trai ning across CPU and GPU architectures\, and how general-purpose (often ref erred to as "foundational") force fields can kick-start data acquisition f or diverse molecular and materials systems.\nThe webinar will demonstrate how aims-PAX reduces the number of required reference calculations by up t o three orders of magnitude while enabling a 10-fold speedup in active lea rning time through optimized resource utilization. It will also demonstrat e how the framework can generate data across gas-phase\, solvated\, and pe riodic systems\, from highly flexible peptides and multiple organic molecu les to explicitly solvated molecules and materials such as perovskites. Ad ditionally\, it will be shown how the automatically collected data sets ar e used to train state-of-the-art MLFF architectures such as MACE\, making aims-PAX a versatile tool for data-driven force-field generation in both a cademic and industrial settings.\nWhat will be covered?\n\n\nDiscover how to automatically generate high-quality reference data sets for machine lea rning force fields using parallel active exploration with aims-PAX and FHI -aims.\n\n\nUnderstand how efficient CPU/GPU workload management substanti ally reduces the overall cost of data generation.\n\n\nParticipate in a gu ided hands-on demonstration on cloud resources to learn how to run these w orkflows with aims-PAX and FHI-aims.\n\n\nBoth the webinar and hands-on se ssion are free to attend without registration. The hands-on session addres ses novice users of FHI-aims\, but may also be interesting to experienced users. Following along and running the calculations in the AWS cloud is c ompletely free\, allowing you to gain practical experience with FHI-aims.  \nWe provide resources for running the calculations during the hands-on session\, but you need to register in advance to obtain access to the comp ute resources.\nRegister now!\nSpeaker:\naims-PAX: Automated Data Generati on through Parallel Active Exploration with FHI-aims\nTobias Henkes\nDocto ral Researcher\, University of Luxembourg\, Luxembourg\n\nWhen? \n\n\n\n\ nWebinar Talks + Q&A\nWednesday\, 07.10.2026:\n08:00 - 09:30 EDT (USA East )\n14:00 - 15:30 CEST (Europe)\n17:30 - 19:00 IST (India)\n \n\n\nHands-O n and Discussions Session 1\nThursday\, 08.10.2026:\n09:00 - 10:30 CEST (E urope)\n12:30 - 14:00 IST (India)\n15:00 - 16:30 CST (China)\n16:00 - 17:3 0 JST (Japan)\n\n\nHands-On and Discussions Session 2\nThursday\, 08.10.20 26:\n07:00 - 08:30 PDT (USA Pacific)\n10:00 - 11:30 EDT (USA East)\n16:00 - 17:30 CEST (Europe)\n \n\n\n\n\n\nPlease choose the hands-on and discus sion block that fits your schedule the best. The hands-on session will as sume a basic understanding of the FHI-aims input files. If you don't have experience with FHI-aims\, you can learn the first steps from this short Y ouTube video:The Basics of FHI-aims\n\nhttps://events.ms1p.org/event/23/ LOCATION:Zoom: https://us06web.zoom.us/j/84093695453 URL:https://events.ms1p.org/event/23/ END:VEVENT END:VCALENDAR