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:20240626T180033Z
LOCATION:Level 2 Lobby
DTSTART;TZID=America/Los_Angeles:20240626T180000
DTEND;TZID=America/Los_Angeles:20240626T190000
UID:dac_DAC 2024_sess237_RESEARCH1873@linklings.com
SUMMARY:Defending Membership Inference Attack on Edge using Trusted Execut
 ion Environments
DESCRIPTION:Work-in-Progress Poster\n\nCheng-Yun Yang, Gowri Ramshankar, S
 udarshan Nambiar, Evan Miller, Xun Zhang, Nicholas Eliopoulos, Purvish Jaj
 al, Xiaolong Wu, and Dave Tian (Purdue University); Shuo-Han Chen (Nationa
 l Yang Ming Chiao Tung University); Chiy-Ferng Perng (Wistron); and Yung-H
 siang Lu (Purdue University)\n\nPrivacy concerns arise from malicious atta
 cks on Deep Neural Network (DNN) applications during sensitive data infere
 nce on edge devices. Our proposed defense method addresses limitations in 
 existing Trusted Execution Environments (TEEs) by employing depth-wise lay
 er partitioning for large DNNs and a model quantization strategy. This enh
 ances protection against both white-box and black-box Membership Inference
  Attacks (MIAs) while accelerating computation. Experiments on Raspberry P
 i 3B+ demonstrate significant reductions in white-box MIA accuracy (up to 
 35.3%) and black-box MIA accuracy (up to 29.6%) for popular DNN models (Al
 exNet, VGG-16, ResNet-20) on CIFAR-100 dataset.\n\nTopic: AI, Autonomous S
 ystems, Cloud, Design, EDA, Embedded Systems, IP, Security
END:VEVENT
END:VCALENDAR