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:3000\, Level 3
DTSTART;TZID=America/Los_Angeles:20250625T111500
DTEND;TZID=America/Los_Angeles:20250625T113000
UID:dac_DAC 2025_sess106_RESEARCH2446@linklings.com
SUMMARY:InsightAlign: A Transferable Physical Design Recipe Recommender Ba
 sed on Design Insights
DESCRIPTION:Hao-Hsiang Hsiao (Georgia Institute of Technology); Sudipto Ku
 ndu, Wei Zeng, Wei-Ting Chan, and Deyuan Guo (Synopsys); and Sung Kyu Lim 
 (Georgia Institute of Technology)\n\nPhysical design tools have complex wo
 rkflows with many different ways of optimizing power, performance and area
  (PPA) out of a large number of options and hyperparameters in different e
 ngines and functionalities. Black box optimization techniques are widely a
 dapted to automate quality-of-result (QoR) exploration. Such exploration o
 ften proves impractical in real-world customer environments due to high co
 mputational demands, lengthy exploration cycles, and the need for large pa
 rallel jobs. To reduce the exploration space for viable compute resource r
 equirement, we propose a novel design methodology to enable transferable l
 earning by incorporating design insights crafted on top of physical design
  experts' experience, and streamlining QoR exploration as a sequence gener
 ation task for best recipe selection, where we apply language model-inspir
 ed alignment techniques to learn the ranking of different recipe sets, ena
 bling our model to generalize beyond known-good manually tuned expert desi
 gn recipes. Extensive evaluations demonstrate our method's superior QoRs a
 nd runtime performance on unseen industrial designs and rigorous benchmark
 s.\n\nTopics: AI\n\nTracks: AI2: AI/ML Application and Infrastructure\n\nS
 ession Chairs: Jun Shiomi (The University of Osaka) and Subhajit Dutta Cho
 wdhury (Advanced Micro Devices (AMD))\n\n
END:VEVENT
END:VCALENDAR