BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
DTSTART:20260308T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251102T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260218T193525Z
UID:C9BF53F2-6191-4C2C-A240-4E84FBCC885B
DTSTART;TZID=America/New_York:20260224T100000
DTEND;TZID=America/New_York:20260224T110000
DESCRIPTION:Physical reservoir computing has been recently proposed as an a
 nalog computational technique that exploits device nonlinearity and crosst
 alk instead of minimizing them\, as in digital logic. As opposed to conven
 tional neural networks\, the reservoir in a reservoir computer is untraine
 d and transforms low-dimensional input information into a significantly-hi
 gher-dimensional representation\, ideally transforming a non-linear tempor
 al classification problem into a linearly separable problem in the high-di
 mensional reservoir state space. A trained linear output layer is therefor
 e used to convert the reservoir state into the classification result.\n\nW
 e proposed and analyzed the dynamics of a reservoir of tightly packed nano
 magnets\, passively relaxing in the presence of their neighbors' stray mag
 netic fields. Micromagnetic simulations demonstrate good expressivity on n
 on-linear\, temporal tasks\, and resource estimations of the electrical MT
 J readout indicate a combined area\, energy\, and delay decrease of 10\,00
 0\,000x compared with CMOS\, necessitating further study.\n\nWe anticipate
  testing the concept using a commercial MRAM array. We have developed a st
 reamlined high-speed FPGA-based interface for the MRAM\, enabling full con
 trol of the MRAM signals. However\, as MRAM MTJs are designed for minimal 
 crosstalk\, it is expected that delayering and deposition techniques will 
 also be needed to increase interactions between devices. We hope that with
  such techniques\, we can demonstrate both the functionality and efficacy 
 of nanomagnet reservoir computing for next-generation neuromorphic applica
 tions.\n\nEdwards\, A.J.\, Bhattacharya\, D.\, Zhou\, P. et al. Passive fr
 ustrated nanomagnet reservoir computing. Commun Phys 6\, 215 (2023). https
 ://doi.org/10.1038/s42005-023-01324-8\n\nSpeaker(s): Alex\n\nRoom: A366\, 
 Bldg: 221\, 100 Bureau Drive\, GAITHERSBURG\, Maryland\, United States\, 2
 0878\, Virtual: https://events.vtools.ieee.org/m/540388
LOCATION:Room: A366\, Bldg: 221\, 100 Bureau Drive\, GAITHERSBURG\, Marylan
 d\, United States\, 20878\, Virtual: https://events.vtools.ieee.org/m/5403
 88
ORGANIZER:daniel.gopman@ieee.org
SEQUENCE:9
SUMMARY:Nanomagnet Reservoir Computing Hardware Platform
URL;VALUE=URI:https://events.vtools.ieee.org/m/540388
X-ALT-DESC:Description: <br /><p class="MsoNormal">Physical reservoir compu
 ting has been recently proposed as an analog computational technique that 
 exploits device nonlinearity and crosstalk instead of minimizing them\, as
  in digital logic. As opposed to conventional neural networks\, the reserv
 oir in a reservoir computer is untrained and transforms low-dimensional in
 put information into a significantly-higher-dimensional representation\, i
 deally transforming a non-linear temporal classification problem into a li
 nearly separable problem in the high-dimensional reservoir state space.&nb
 sp\; A trained linear output layer is therefore used to convert the reserv
 oir state into the classification result.</p>\n<p class="MsoNormal">We pro
 posed and analyzed the dynamics of a reservoir of tightly packed nanomagne
 ts\, passively relaxing in the presence of their neighbors' stray magnetic
  fields.&nbsp\; Micromagnetic simulations demonstrate good expressivity on
  non-linear\, temporal tasks\, and resource estimations of the electrical 
 MTJ readout indicate a combined area\, energy\, and delay decrease of 10\,
 000\,000x compared with CMOS\, necessitating further study.</p>\n<p class=
 "MsoNormal">We anticipate testing the concept using a commercial MRAM arra
 y.&nbsp\; We have developed a streamlined high-speed FPGA-based interface 
 for the MRAM\, enabling full control of the MRAM signals.&nbsp\; However\,
  as MRAM MTJs are designed for minimal crosstalk\, it is expected that del
 ayering and deposition techniques will also be needed to increase interact
 ions between devices.&nbsp\; We hope that with such techniques\, we can de
 monstrate both the functionality and efficacy of nanomagnet reservoir comp
 uting for next-generation neuromorphic applications.</p>\n<p class="MsoNor
 mal">Edwards\, A.J.\, Bhattacharya\, D.\, Zhou\, P. et al. Passive frustra
 ted nanomagnet reservoir computing. Commun Phys 6\, 215 (2023).&nbsp\;<a h
 ref="https://doi.org/10.1038/s42005-023-01324-8" target="_blank" rel="noop
 ener">https://doi.org/10.1038/s42005-023-01324-8</a></p>
END:VEVENT
END:VCALENDAR