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:20230731T004029Z
UID:5B14BC48-0248-49BB-A600-86080109E784
DTSTART;TZID=America/New_York:20230727T100000
DTEND;TZID=America/New_York:20230727T120000
DESCRIPTION:Abstract\n\nMachine learning (ML) has recently been applied for
  the classification of radio frequency (RF) signals. One use case of inter
 est relates to the discernment between different wireless protocols that o
 perate over a shared and potentially contested spectrum. Although highly a
 ccurate classifiers have been developed for various wireless scenarios\, r
 esearch points to the vulnerability of such classifiers to adversarial mac
 hine learning (AML) attacks. In one such attack\, a surrogate deep neural 
 network (DNN) model is trained by the attacker to produce intelligently cr
 afted low power “perturbations” that degrade the classification accura
 cy of the legitimate classifier. In this talk\, I will first present sever
 al novel DNN protocol classifiers that we designed for a shared spectrum e
 nvironment. These classifiers performed quite well in both simulations and
  OTA experimentation\, considering benign (non-adversarial) noise. I will 
 then present several AML techniques that an attacker may use to generate l
 ow power perturbations. When combined with a legitimate signal\, these per
 turbations are shown to uniformly degrade the classification accuracy\, ev
 en in the very high SNR regime. Different attack models are studied\, depe
 nding on how much information the attacker has about the defender’s clas
 sifier. Finally\, I will discuss possible defense mechanisms as well as ot
 her research efforts related to detection of adversarial transmissions.\n\
 nCo-sponsored by: Dr. Jun Yan\n\nRoom: EV001.162\, Bldg: EV001.162\, 1515 
 St. Catherine St. West\, Montreal\, Quebec H3G 2W1\, Montreal\, Quebec\, C
 anada\,  H3G 2W1
LOCATION:Room: EV001.162\, Bldg: EV001.162\, 1515 St. Catherine St. West\, 
 Montreal\, Quebec H3G 2W1\, Montreal\, Quebec\, Canada\,  H3G 2W1
ORGANIZER:anader.benyamin@ieee.org
SEQUENCE:3
SUMMARY:Adversarial Machine Learning Attacks on RF Signal Classifiers
URL;VALUE=URI:https://events.vtools.ieee.org/m/368001
X-ALT-DESC:Description: <br /><h2>Abstract</h2>\n<p><span class="xlarge-tex
 t">Machine learning (ML) has recently been applied for the classification 
 of radio frequency (RF) signals. One use case of interest relates to the d
 iscernment between different wireless protocols that operate over a shared
  and potentially contested spectrum. Although highly accurate classifiers 
 have been developed for various wireless scenarios\, research points to th
 e vulnerability of such classifiers to adversarial machine learning (AML) 
 attacks. In one such attack\, a surrogate deep neural network (DNN) model 
 is trained by the attacker to produce intelligently crafted low power &ldq
 uo\;perturbations&rdquo\; that degrade the classification accuracy of the 
 legitimate classifier. In this talk\, I will first present several novel D
 NN protocol classifiers that we designed for a shared spectrum environment
 . These classifiers performed quite well in both simulations and OTA exper
 imentation\, considering benign (non-adversarial) noise. I will then prese
 nt several AML techniques that an attacker may use to generate low power p
 erturbations. When combined with a legitimate signal\, these perturbations
  are shown to uniformly degrade the classification accuracy\, even in the 
 very high SNR regime. Different attack models are studied\, depending on h
 ow much information the attacker has about the defender&rsquo\;s classifie
 r. Finally\, I will discuss possible defense mechanisms as well as other r
 esearch efforts related to detection of adversarial transmissions.</span><
 /p>
END:VEVENT
END:VCALENDAR