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:20260730T152638Z
LOCATION:
DTSTART;TZID=America/Los_Angeles:20260728T125600
DTEND;TZID=America/Los_Angeles:20260728T125600
UID:dac_DAC 2026_sess315_RESEARCH1936@linklings.com
SUMMARY:RTFL: Energy-Aware Federated Learning for AIoT Design Via Adaptive
  Quantization-based Multi-Agent Scheduling
DESCRIPTION:Jun Xia (University of Notre Dame) and Junqi Zhang, Zhaorong Z
 hu, Wenjie Chen, and Mingsong Chen (East China Normal University)\n\nAltho
 ugh Federated Learning (FL) is becoming increasingly popular in designing 
 Artificial Intelligence of Things (AIoT) applications, due to the varying 
 computing and communication capabilities of resource-constrained devices, 
 it suffers from the problems of slow convergence and poor training perform
 ance, especially when devices are powered by batteries. To address these i
 ssues, this paper introduces RTFL, a novel energy-aware Real-Time Federate
 d Learning framework based on Multi-Agent Reinforcement Learning (MARL),  
 aiming to enhance the  knowledge sharing across AIoT devices within a spec
 ified training time constraint. Specifically, RTFL employs an Adaptive Qua
 ntization-based Multi-Agent Scheduling (AQMAS) strategy, enabling a team o
 f agents to intelligently select devices with specific model quantization 
 levels for each round of local training, taking into account the resource 
 constraints (e.g., remaining battery power, computing capability, and comm
 unication bandwidth) of the current devices. By facilitating collaboration
  among agents through reinforcement learning, our approach enables devices
  to maximize their contributions to forming an optimal global model, while
  balancing the trade-off between the accuracy of quantized models and the 
 limited resources available on each device. Comprehensive experiments show
  that RTFL not only accelerates the convergence of FL training, but also e
 ncourages devices to participate in more rounds of knowledge aggregation, 
 thereby significantly improving overall training performance.\n\nTopics: Q
 uantum\n\nTrack: DES6. Quantum Computing, Systems\n\nSession Chairs: Hokeu
 n Kim (Arizona State University); BaekGyu Kim (Daegu Gyeongbuk Institute o
 f Science and Technology); Xiang Chen (Peking University); Sangyeob Kim (Y
 onsei University); Robert M. Radway (University of Pennsylvania); Prashant
 h Mohan (University of Pittsburgh); Xiaoxuan Yang (University of Virginia,
  Stanford University); Ramtin Zand (University of South Carolina); Hadi Ka
 mali (University of Central Florida); Jiafeng Xie (Villanova University); 
 Prabuddha Chakraborty (University of Maine); Junichiro Kadomoto (Universit
 y of Tokyo); and Yosuke Ueno (RIKEN)\n\n
END:VEVENT
END:VCALENDAR