BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:HPC v praksi III: RAG – združevanje znanja in umetne inteligenc
 e
DTSTART:20261106T080000Z
DTEND:20261106T160000Z
DTSTAMP:20260928T002400Z
UID:indico-event-3824@indico.ijs.si
CONTACT:events@slaif.si
DESCRIPTION:Speakers: Pavle Boškoski\, Srdjan Šrbić\, Robi Podtržnik\,
  Biljana Mileva Boshkoska\n\nCourse provider: Faculty of Information Studi
 es in Novo mesto (FIŠ)Instructors: Biljana Mileva Boshkoska (FIŠ)\, Srdj
 an Šrbić (FIŠ)\, Robi Podtržnik (FIŠ)\, Pavle Boškoski (FIŠ)\nThis 
 intensive course is designed for AI developers and researchers seeking to 
 overcome the limitations of individual processing units in demanding compu
 tational projects. The focus of the course is on the practical implementat
 ion of massive parallel processing using multiple graphics cards (Multi-GP
 U). Participants will learn how to use advanced distributed computing stra
 tegies to drastically accelerate real-world algorithms that are key to mod
 ern artificial intelligence and probabilistic modeling.\nLearning objectiv
 es: The primary goal of the course is to equip participants with the skill
 s to implement the Distributed Data Parallel (DDP) strategy as a fundament
 al tool for parallelizing complex computational tasks. Rather than focusin
 g solely on the theoretical aspects of models\, we will concentrate on how
  to effectively use DDP to scale Monte Carlo simulations and Variational B
 ayes inference across multiple nodes. Participants will learn to optimize 
 data synchronization and manage distributed tensors on high-performance in
 frastructure\, enabling the transfer of deep learning methodologies to the
  field of advanced statistical analysis.\nCourse content: The course conte
 nt begins with the technical configuration of a DDP environment for Multi-
 GPU systems\, focusing on process orchestration and communication protocol
 s between GPU units. In the core part of the course\, participants impleme
 nt parallel versions of Monte Carlo algorithms\, where DDP serves to distr
 ibute massive sampling tasks across the entire HPC cluster. This is follow
 ed by a practical module on parallelizing Variational Bayes inference\, wh
 ere we will use DDP to accelerate optimization steps in approximating dist
 ributions over large datasets. The course concludes with code optimization
  for GPU acceleration\, allowing participants to immediately apply these t
 echniques to their own AI projects.\nLearning outcomes: Upon completion of
  the course\, participants will be able to independently configure and use
  the DDP protocol for the parallel execution of any AI and statistical alg
 orithms on Multi-GPU infrastructure. They will gain practical knowledge in
  converting serial Monte Carlo simulations into highly scalable distribute
 d processes. They will be trained to use DDP in variational methods\, enab
 ling significantly faster training of complex probabilistic models. With t
 hese skills\, researchers will be prepared to tackle the most demanding co
 mputational challenges where the use of HPC resources is necessary to achi
 eve results within a reasonable timeframe.\n\nhttps://indico.ijs.si/event/
 3824/
IMAGE;VALUE=URI:https://indico.ijs.si/event/3824/logo-1596601024.png
LOCATION:Slovenija
URL:https://indico.ijs.si/event/3824/
END:VEVENT
END:VCALENDAR