BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20211208T164413Z
LOCATION:Level 2 - Lobby
DTSTART;TZID=America/Los_Angeles:20211208T180000
DTEND;TZID=America/Los_Angeles:20211208T190000
UID:dac_DAC 2021_sess243_WIP206@linklings.com
SUMMARY:Securing Deep Neural Networks Against Adversarial Attacks through 
 Voltage Overscaling
DESCRIPTION:Networking Reception, Work-in-Progress Poster\n\nSecuring Deep
  Neural Networks Against Adversarial Attacks through Voltage Overscaling\n
 \nIslam\n\nDeep neural networks (DNNs) are shown to be vulnerable to adver
 sarial attacks. Previously proposed defenses against adversarial attacks r
 equire substantial additional overheads, making it challenging to deploy t
 hese solutions in devices under power and computational resource constrain
 ts, such as embedded systems and the Edge.  In this paper, we explore the 
 use of voltage over-scaling (VOS) as a lightweight defense against adversa
 rial attacks. Specifically, we exploit the stochastic timing violations of
  VOS to implement a moving-target defense for DNNs. Our experimental resul
 ts demonstrate that VOS offers effective defense, does not require any sof
 tware/hardware modifications, and offers a reduction in power consumption.
 \n\nVirtual Program: Presented In-Person
END:VEVENT
END:VCALENDAR