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:20211208T164408Z
LOCATION:3018
DTSTART;TZID=America/Los_Angeles:20211207T143500
DTEND;TZID=America/Los_Angeles:20211207T150000
UID:dac_DAC 2021_sess157_RESEARCH1249@linklings.com
SUMMARY:A3C-S: Automated Agent Accelerator Co-Search towards Efficient Dee
 p Reinforcement Learning
DESCRIPTION:Research Manuscript\n\nA3C-S: Automated Agent Accelerator Co-S
 earch towards Efficient Deep Reinforcement Learning\n\nZhang\n\nDriven by 
 the explosive interest in applying deep reinforcement learning (DRL) agent
 s for numerous real-time control and decision-making applications, there h
 as been a growing demand to deploy DRL agents to empower daily-life intell
 igent devices, while the prohibitive complexity of DRL stands at the odds 
 with the limited on-device resources. In this work, we propose an Automate
 d Agent Accelerator Co-Search (A3C-S) framework, which to our best knowled
 ge is the first to automatically co-search the optimally matched DRL agent
 s and accelerators that maximize both the test scores and hardware efficie
 ncy. Extensive experiments consistently validate the superiority of our A3
 C-S over state-of-the-art techniques.\n\nTopic: Design\n\nKeyword: AI/ML S
 ystem Design\n\nVirtual Program: Presented In-Person
END:VEVENT
END:VCALENDAR