BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
DTSTART:20230312T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20231105T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20231130T170811Z
UID:8ADF5068-CCBA-4C68-B4CC-8D90D2E4DB24
DTSTART;TZID=America/New_York:20230707T150000
DTEND;TZID=America/New_York:20230707T163000
DESCRIPTION:A recent report by the US Department of Energy defines the area
  of scientific machine learning as “a core component of artificial intel
 ligence (AI) and a computational technology that can be trained\, with sci
 entific data\, to augment or automate human skills”\, which has “the\n
 potential to transform science and energy research”.\n\nWe explore the p
 otential of scientific machine learning methods to problems in computation
 al electromagnetics starting\nfrom standard microwave structure design and
  multiphysics modeling\, employing an unsupervised learning strategy based
  on Physics-Informed Neural Networks (PINN). PINNs directly integrate phys
 ical laws into their loss function\, so that the training process does not
  rely on the generation of ground truth data from a large number of simula
 tions (as in typical neural networks).\n\nMoreover\, we demonstrate the im
 pact of machine learning on the computational modeling of radiowave propag
 ation scenarios. We build convolutional neural network models that can pro
 cess the geometry of indoor environments\, along with physics-inspired par
 ameters\, to rapidly estimate received signal strength (RSS) maps. We show
  the *generalizability* of these models\, which is their ability to "learn
 " the physics of radiowave propagation\n\nand produce accurate modeling pr
 edictions in new geometries well beyond those included in their training s
 et. These models can be used to rapidly optimize the position of transmitt
 ers in wireless area networks\, to maximize\ncoverage or other relevant me
 trics.\n\nCo-sponsored by: STARaCom\n\nSpeaker(s): Costas Sarris\n\nRoom: 
 Room 603\, Bldg: McConnell Eng. Building \, McGill Unversity\, Montreal\, 
 Quebec\, Canada
LOCATION:Room: Room 603\, Bldg: McConnell Eng. Building \, McGill Unversity
 \, Montreal\, Quebec\, Canada
ORGANIZER:roni.khazaka@mcgill.ca
SEQUENCE:37
SUMMARY:Scientific Machine Learning for Computational electromagnetics: fro
 m Microwave Circuits to Radiowave Propagation
URL;VALUE=URI:https://events.vtools.ieee.org/m/365732
X-ALT-DESC:Description: <br /><p>A recent report by the US Department of En
 ergy defines the area of scientific machine learning as &ldquo\;a core com
 ponent of artificial intelligence (AI) and a computational technology that
  can be trained\, with scientific data\, to augment or automate human skil
 ls&rdquo\;\, which has &ldquo\;the<br />potential to transform science and
  energy research&rdquo\;.<br /><br />We explore the potential of scientifi
 c machine learning methods to problems in computational electromagnetics s
 tarting<br />from standard microwave structure design and multiphysics mod
 eling\, employing an unsupervised learning strategy based on Physics-Infor
 med Neural Networks (PINN). PINNs directly integrate physical laws into th
 eir loss function\, so that the training process does not rely on the gene
 ration of ground truth data from a large number of simulations (as in typi
 cal neural networks).<br /><br />Moreover\, we demonstrate the impact of m
 achine learning on the computational modeling of radiowave propagation sce
 narios. We build convolutional neural network models that can process the 
 geometry of indoor environments\, along with physics-inspired parameters\,
  to rapidly estimate received signal strength (RSS) maps. We show the *gen
 eralizability* of these models\, which is their ability to "learn" the phy
 sics of radiowave propagation</p>\n<p>and produce accurate modeling predic
 tions in new geometries well beyond those included in their training set. 
 These models can be used to rapidly optimize the position of transmitters 
 in wireless area networks\, to maximize<br />coverage or other relevant me
 trics.</p>
END:VEVENT
END:VCALENDAR