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:20260730T152640Z
LOCATION:
DTSTART;TZID=America/Los_Angeles:20260728T173000
DTEND;TZID=America/Los_Angeles:20260728T173000
UID:dac_DAC 2026_sess317_RESEARCH1922@linklings.com
SUMMARY:Pegasus: Personalized and Communication-Efficient Mixture-of-Exper
 ts System for Distributed Edge LLM Inference
DESCRIPTION:Zhixuan Liao and Kongyange Zhao (Sun Yat-sen University), Tao 
 Ouyang (Central South University), Liekang Zeng (The Chinese University of
  Hong Kong), and Siqi Luo and Xu Chen (Sun Yat-sen University)\n\nLarge La
 nguage Models (LLMs) are increasingly pushed to the edge to reduce latency
 , yet limited memory and compute make full-model deployment infeasible. Mi
 xture-of-Experts (MoE) offers a lightweight alternative, but edge nodes ca
 n host only a few experts, creating persistent mismatches between dynamic 
 query semantics and locally available expert models under resource constra
 ints. We present Pegasus, a collaborative MoE inference system for multi-e
 dge deployments that leverages the spatiotemporal correlations inherent in
  distributed query semantics. Pegasus integrates three innovations: (i) a 
 similarity-based expert deployment mechanism using an efficient heuristic 
 metric to assess the semantic relevance between queries and experts across
  nodes and over time; (ii) a personalized gating design that selectively f
 ine-tunes a subset of gating parameters on each node to balance expert acc
 uracy and communication latency; (iii) an intra-node online scheduling alg
 orithm with adaptive batching for efficient memory utilization. Extensive 
 performance evaluations corroborate that Pegasus achieves 11.8x lower infe
 rence latency than state-of-the-art distributed MoE frameworks, demonstrat
 ing high-throughput and communication-efficient edge LLM inference.\n\nTra
 ck: EDA7-II. Physical Design and Verification, Design\n\nSession Chairs: C
 hristian Pilato (Politecnico di Milano); Mohsen Imani (University of Calif
 ornia, Irvine); Flavio Ponzina (San Diego State University); Amin Farshidi
  (Cadence Design Systems, Inc.); Yonggan Fu (Georgia Institute of Technolo
 gy, NVIDIA); Dimitrios Tsaras (Huawei, Hong Kong University of Science and
  Technology (HKUST)); Anh Nguyen (Fujitsu Limited); Ruiyang Qin (Villanova
  University); Jiliang Zhang (Hunan University); Meng Li (Peking University
 ); and Rassul Bairamkulov (Advanced Micro Devices (AMD))\n\n
END:VEVENT
END:VCALENDAR