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:20260724T151409Z
LOCATION:Bldg. 6 - 004
DTSTART;TZID=Europe/Stockholm:20260629T160000
DTEND;TZID=Europe/Stockholm:20260629T180000
UID:submissions.pasc-conference.org_PASC26_sess159@linklings.com
SUMMARY:MS2C - Serving Inference: Leveraging HPCs in the Age of Generative
  AI
DESCRIPTION:Access the recording\n\nOrganizer(s): Tobias Hodel (University
  of Bern, Switzerland), and Sukanya Nath (University of Bern)\n\nGenerativ
 e AI is reshaping how scientific outputs are produced, yet most widely use
 d tools are operated by commercial providers whose practices around proces
 sing, generating, and storing user inputs are often unclear. This lack of 
 transparency raises serious concerns for research organizations handling s
 ensitive or regulated data and complicates the responsible use of even non
 -regulated scientific content. At the same time, many commercially served 
 models are closed-source and subject to frequent, opaque updates, limiting
  reproducibility and undermining alignment with FAIR principles. As open-w
 eight and open-source models proliferate, HPC infrastructures are emerging
  as a promising alternative for hosting and providing controlled access to
  AI within research environments. However, most HPC systems were not desig
 ned for continuous, GPU-based inference services, creating technical and o
 perational challenges spanning deployment, scheduling, reliability, user a
 ccess, and governance. Consequently, institutions are developing ad hoc so
 lutions with limited opportunities to exchange patterns and lessons learne
 d. This minisymposium convenes a panel of institutions actively building s
 uch capabilities to compare approaches and discuss best practices, minimal
  viable solutions, and ideal configurations for serving generative AI on H
 PC. To broaden participation, we will use a short questionnaire to structu
 re contributions across four dimensions: technical setup, usage policies, 
 documentation practices, and monitoring/oversight mechanisms.\n\nA Human-i
 n-the-Loop Scoping Review Screening Pipeline Using Self-Hosted Large Langu
 age Models: An Example in Sport Science\n\nScoping reviews are labor-inten
 sive efforts to screen and extract paper data. Large Language Models (LLMs
 ) seem to offer efficiency gains. But to address data privacy and reproduc
 ibility issues with cloud-based LLMs, and adhering to JBI/PRISMA-ScR guide
 lines, we present a human-in-the-loop review pi...\n\n\nKai Michael Gensit
 z (University of Bern); Shawan Mohammed (RWTH Aachen University); Daniela 
 E. Ströckl (Carinthia University of Applied Sciences); Marc Augustin (Prot
 estant University of Applied Sciences, Bochum); Claudio R. Nigg (Universit
 y of Bern); and Ciara McCormack (National University of Ireland, Maynooth)
 \n---------------------\nLLM Infrastructure on HPC: Workflows, Constraints
 , and Solutions\n\nThe integration of Large Language Models (LLMs) into ac
 ademic research is severely constrained by data privacy regulations. Resea
 rchers handling sensitive, GDPR-protected data cannot utilize commercial c
 loud APIs, necessitating the local deployment of open-weight LLMs on High-
 Performance Computing (...\n\n\nAhmad Alhineidi (University of Bern, Data 
 Science Lab)\n---------------------\nUNIBE's GPUStack as an Example: Repor
 ting on User Needs\n\nThe University of Bern has developed a proof-of-conc
 ept platform using GPUstack technology, providing a secure, institutionall
 y controlled environment for deploying generative AI models. Designed for 
 researchers working outside commercial cloud infrastructures, it enables d
 irect management of comput...\n\n\nTobias Hodel (University of Bern)\n----
 -----------------\nPanel Discussion on Serving Generative AI on HPC\n\nThi
 s session will be a panel discussion about ad hoc solutions of serving gen
 erative AI models over HPC  infrastructure including best practices, minim
 al viable solutions, and ideal configurations in research contexts.\n\n\nS
 ukanya Nath (University of Bern, Data Science Lab)\n\nDomain: Chemistry an
 d Materials, Climate, Weather, and Earth Sciences, Applied Social Sciences
  and Humanities, Engineering, Life Sciences, Physics, Computational Method
 s and Applied Mathematics\n\nSession Chairs: Tobias Hodel (University of B
 ern, Switzerland) and Sukanya Nath (University of Bern, Data Science Lab)
END:VEVENT
END:VCALENDAR