BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
DTSTART:20240310T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20241103T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240915T190102Z
UID:D12E5154-BBD9-42F1-906D-F46D4ECD5818
DTSTART;TZID=America/New_York:20240912T170000
DTEND;TZID=America/New_York:20240912T190000
DESCRIPTION:This talk will provide an overview of deep learning application
 s in nanophotonic device design\, focusing on generative neural networks. 
 To achieve inverse design in nanophotonics\, optimization of tens of thous
 ands of 'pixels' is typically required. The adjoint method\, a popular loc
 al optimization approach\, often necessitates multiple optimization runs. 
 Generative deep learning builds on existing data to generate new designs w
 ith specified target specifications such as transmission/reflection spectr
 a. For instance\, datasets optimized for discrete wavelengths (e.g.\, wave
 length splitters) or splitting ratios (e.g.\, power splitters) can be used
  to generate devices with arbitrary wavelength or splitting ratios. We dem
 onstrate examples using conditional variational autoencoders (CVAE) and de
 noising diffusion probabilistic models (DDPM) for applications in planar w
 aveguide devices\, metasurface gratings\, and plasmonic gratings. Addition
 ally\, we introduce the concept of latent space optimization and transfer 
 learning.\n\nSpeaker(s): Keisuke\n\nAgenda: \n5:00 - 5:30 PM Assembly and 
 buffet dinner\n\n5:30 - 6:30 PM Presentation\n\n6:00 - 7:00 PM Networking\
 n\nRoom: 6A-106\, Nokia Bell Labs\, 600 Mountain Ave\, Murray Hill\, New J
 ersey\, United States\, 07974
LOCATION:Room: 6A-106\, Nokia Bell Labs\, 600 Mountain Ave\, Murray Hill\, 
 New Jersey\, United States\, 07974
ORGANIZER:nchand@ieee.org
SEQUENCE:41
SUMMARY:Machine Learning and Photonic Devices
URL;VALUE=URI:https://events.vtools.ieee.org/m/430961
X-ALT-DESC:Description: <br /><p class="MsoNormal" style="text-align: justi
 fy\; text-justify: inter-ideograph\;">This talk will provide an overview o
 f deep learning applications in nanophotonic device design\, focusing on g
 enerative neural networks. To achieve inverse design in nanophotonics\, op
 timization of tens of thousands of 'pixels' is typically required. The adj
 oint method\, a popular local optimization approach\, often necessitates m
 ultiple optimization runs. Generative deep learning builds on existing dat
 a to generate new designs with specified target specifications such as tra
 nsmission/reflection spectra. For instance\, datasets optimized for discre
 te wavelengths (e.g.\, wavelength splitters) or splitting ratios (e.g.\, p
 ower splitters) can be used to generate devices with arbitrary wavelength 
 or splitting ratios. We demonstrate examples using conditional variational
  autoencoders (CVAE) and denoising diffusion probabilistic models (DDPM) f
 or applications in planar waveguide devices\, metasurface gratings\, and p
 lasmonic gratings. Additionally\, we introduce the concept of latent space
  optimization and transfer learning.</p><br /><br />Agenda: <br /><p>5:00 
 - 5:30 PM&nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\;Assembly and buffe
 t dinner</p>\n<p>5:30 - 6:30 PM&nbsp\; &nbsp\; &nbsp\; &nbsp\; &nbsp\;&nbs
 p\; Presentation</p>\n<p>6:00 - 7:00 PM&nbsp\; &nbsp\; &nbsp\; &nbsp\; &nb
 sp\;&nbsp\; Networking</p>
END:VEVENT
END:VCALENDAR