BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:PDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0700 TZOFFSETTO:-0800 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:PST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20261008T021148Z UID:A255F4D3-FA88-4968-A378-BF0B527B0FC6 DTSTART;TZID=America/Los_Angeles:20261030T093000 DTEND;TZID=America/Los_Angeles:20261030T103000 DESCRIPTION:Abstract:\n\nA recent report by the US Department of Energy def ines the area of scientific machine learning as “a core component of art ificial intelligence (AI) and a computational technology that can be train ed\, with scientific data\, to augment or automate human skills”\, which has “the potential to transform science and energy research”. In this presentation\, we discuss the potential of scientific machine learning me thods to problems in computational electromagnetics starting from standard electromagnetic structure analysis and multi-physics modeling in the time -domain\, employing an unsupervised learning strategy based on Physics-Inf ormed Neural Networks (PINN). PINNs directly integrate physical laws into their loss function\, so that the training process does not rely on the ge neration of training ground truth data from simulations. We demonstrate th e unconditionally stable solution of coupled electromagnetic-thermal probl ems based on PINNs and introduce a physics-informed deep-operator-network (PI-DON) for the unconditionally stable solution of Maxwell’s equations. We show applications of the PI-DON to the analysis\, design and uncertain ty quantification study of finite and defect periodic structures\, includi ng frequency selective surfaces and metasurfaces\, demonstrating significa nt computational efficiency gains with respect to the conventional finite- difference time-domain technique (FDTD)\, including training time of the n eural network.\n\nMoreover\, we demonstrate the impact of machine learning on the computational modeling of radiowave propagation scenarios. We buil d convolutional neural network models that can process the geometry of ind oor environments\, along with physics-inspired parameters\, to rapidly est imate received signal strength (RSS) maps. Emphasis is placed on the gener alizability of these models\, which is their ability to "learn" the physic s of radiowave propagation and produce accurate modeling predictions in ne w geometries well beyond those included in their training set.\n\nSpeaker( s): Costas Sarris\n\nVirtual: https://events.vtools.ieee.org/m/583371 LOCATION:Virtual: https://events.vtools.ieee.org/m/583371 ORGANIZER:vgadde@ieee.org SEQUENCE:34 SUMMARY:Scientific machine learning for electromagnetic field computations URL;VALUE=URI:https://events.vtools.ieee.org/m/583371 X-ALT-DESC:Description: <br /><p><strong>Abstract:</strong></p>\n<p class=" MsoNormal">A recent report by the US Department of Energy defines the area of scientific machine learning as&nbsp\;&ldquo\;a core component of artif icial intelligence (AI) and a computational technology that can be trained \, with scientific data\, to augment or automate human skills&rdquo\;\, wh ich has &ldquo\;the potential to transform science and energy research&rdq uo\;. In this presentation\, we discuss the potential of scientific machin e learning methods to problems in computational electromagnetics starting from standard electromagnetic structure analysis and multi-physics modelin g in the time-domain\, employing an unsupervised learning strategy based o n Physics-Informed Neural Networks (PINN). PINNs directly integrate physic al laws into their loss function\, so that the training process does not r ely on the generation of training ground truth data from simulations. We d emonstrate the unconditionally stable solution of coupled electromagnetic- thermal problems based on PINNs and introduce a physics-informed deep-oper ator-network (PI-DON) for the unconditionally stable solution of Maxwell&r squo\;s equations. We show applications of the PI-DON to the analysis\, de sign and uncertainty quantification study of finite and defect periodic st ructures\, including frequency selective surfaces and metasurfaces\, demon strating significant computational efficiency gains with respect to the co nventional finite-difference time-domain technique (FDTD)\, <em>including training time of the neural network.</em></p>\n<p class="MsoNormal">Moreov er\, we demonstrate the impact of machine learning on the computational mo deling of radiowave propagation scenarios. We build convolutional neural n etwork models that can process the geometry of indoor environments\, along with physics-inspired parameters\, to rapidly estimate received signal st rength (RSS) maps. Emphasis is placed on the <strong>generalizability</str ong> of these models\, which is their ability to "learn" the physics of ra diowave propagation and produce accurate modeling predictions in new geome tries well beyond those included in their training set.</p>\n<p>&nbsp\;</p > END:VEVENT END:VCALENDAR