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:20260402T024534Z
LOCATION:3002\, Level 3
DTSTART;TZID=America/Los_Angeles:20250623T111500
DTEND;TZID=America/Los_Angeles:20250623T113000
UID:dac_DAC 2025_sess120_RESEARCH718@linklings.com
SUMMARY:PUFiM: A Robust and Efficient FeFET-Based Security Solution Mergin
 g Physical Unclonable Function with Compute-in-Memory for Edge AI
DESCRIPTION:Taixin Li (Tsinghua University), Thomas Kämpfe (Fraunhofer IPM
 S), Jianfeng Wang (Tsinghua University), Kai Ni (University of Notre Dame)
 , Vijaykrishnan Narayanan (Pennsylvania State University), and Huazhong Ya
 ng and Xueqing Li (Tsinghua University)\n\nCompute-in-memory (CiM) has bec
 ome a promising candidate for edge AI by reducing data movements through i
 n-situ operations. However, this emerging computational paradigm also pose
 s the vulnerability of model leakage as the weights are stored in plaintex
 t for computing. While prior works have explored lightweight encryption me
 thods, CiM is usually considered a separate module instead of a system com
 ponent, leaving the origin of keys unclear and unprotected. Physical unclo
 nable functions (PUFs) offer a potential origin of keys, but a comprehensi
 ve framework for securing key generation and delivery remains lacking. Bes
 ides, the complementary ciphertext storage incurs substantial costs and de
 grades the performance.\n\nThis work proposes PUFiM, a robust and efficien
 t security solution for edge computing based on ferroelectric FETs (FeFETs
 ). For the first time, a strong PUF is synergized with CiM to enable authe
 ntication, key generation, and encrypted computations within a unified arr
 ay for comprehensive protection. To achieve this synergization, a high-den
 sity hybrid storage and computation approach combining PUF and weight bits
  via multi-level cell (MLC) FeFETs is proposed. Besides, two PUF enhanceme
 nt techniques and a novel mapping scheme are developed to improve security
  and efficiency further. Results show that PUFiM withstands PUF modeling a
 ttacks with up to 10M samples. Moreover, PUFiM reduces the inference accur
 acy by >60% under 95% key leakage and achieves >9.7× compute density and >
 1.2× energy efficiency improvement compared with the state-of-the-art SRAM
 /NVM secure CiMs.\n\nTopics: Design\n\nTracks: DES2A: In-memory and Near-m
 emory Computing Circuits\n\nSession Chairs: Yu Cao (University of Minnesot
 a) and Sumitha George (North Dakota State University)\n\n
END:VEVENT
END:VCALENDAR