BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20241120T082409Z
LOCATION:HG F 3
DTSTART;TZID=Europe/Stockholm:20240604T160000
DTEND;TZID=Europe/Stockholm:20240604T180000
UID:submissions.pasc-conference.org_PASC24_sess112@linklings.com
SUMMARY:MS4B - Machine Learning Support for the Lifetime of Software (ML4S
 W)
DESCRIPTION:Minisymposium\n\nScientific simulations running on High Perfor
 mance Computing (HPC) systems play a critical role\nin advancing science a
 nd engineering. The HPC community stands to gain significantly by applying
 \ncutting edge AI technologies, such as Large Language Models (LLMs), Deep
  Neural Networks\n(DNNs), or Transformers, in various aspects of scientifi
 c software development and execution. The Machine Learning Support for The
  Lifetime of Software (ML4SW) minisymposium aims to establish a platform w
 here scientists, developers, and system programmers can come together to e
 xchange ideas and explore how artificial intelligence can help in the effe
 ctive use of future systems as well as how Scientific Machine Learning can
  be scaled on HPC systems.\n\nLarge Language Models for Parallel and HPC C
 ode\n\nLarge Language Model-based coding assistants have already proven to
  be extremely useful tools for aiding the efficiency and correctness of so
 ftware developers. Adapting these tools in scientific software development
  will greatly improve the quality and quantity of scientific code being de
 veloped, le...\n\n\nDaniel Nichols (University of Maryland)\n-------------
 --------\nLearning to Predict and Improve Build Successes in Package Ecosy
 stems Using Graph Neural Networks\n\nModern software has reached an unprec
 edented level of complexity, consisting of tens or even hundreds of depend
 encies on various packages. In order to tackle this complexity, software e
 cosystems rely on automated package managers to analyze compatibility cons
 traints among different packages and sele...\n\n\nHarshitha Menon (Lawrenc
 e Livermore National Laboratory)\n---------------------\nHigh Performance 
 Kernel Code Generation Using Generative AI\n\nGenerative Artificial Intell
 igence (AI) technologies, such as GPT and Llama, have shown promise in fac
 ilitating code generation across a variety of programming languages. Howev
 er, the domain of high-performance scientific computing, which demands spe
 cialized expertise, presents unique challenges tha...\n\n\nPedro Valero-La
 ra, William Godoy, and Keita Teranishi (Oak Ridge National Laboratory); Mu
 stafa Al Lali and Alexis Huante (Texas A&M University); and Prasanna Balap
 rakash and Jeffery Vetter (Oak Ridge National Laboratory)\n---------------
 ------\nMachine Learning for Performance Engineering Across Applications\n
 \nDeveloping fast and portable HPC codes is an evolving process that spans
  the entire lifetime of an application, adapting to changes in target hard
 ware and new software optimization techniques over time. Unfortunately, be
 cause of subtle differences in codes, the lessons learned from optimizing 
 one ap...\n\n\nTal Ben-Nun (Lawrence Livermore National Laboratory)\n\nDom
 ain: Computational Methods and Applied Mathematics\n\nSession Chairs: Flor
 ina Ciorba (University of Basel) and Harshitha Menon (Lawrence Livermore N
 ational Laboratory)
END:VEVENT
END:VCALENDAR