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:20260605T154541Z
LOCATION:Bldg. 8 - Room B 101
DTSTART;TZID=Europe/Stockholm:20260629T143000
DTEND;TZID=Europe/Stockholm:20260629T150000
UID:submissions.pasc-conference.org_PASC26_sess110_msa250@linklings.com
SUMMARY:HPC Infrastructure for Multimodal Medical AI: Lessons from Bridges
 -2 and Neocortex
DESCRIPTION:Paola Buitrago (Pittsburgh Supercomputing Center)\n\nLarge mul
 timodal models (LMMs) are rapidly transforming medical AI by enabling inte
 grated analysis across imaging, clinical text, structured health records, 
 and biomedical data. These models are reshaping how researchers approach d
 iagnostics, clinical decision support, and biomedical discovery by connect
 ing data modalities that have traditionally remained isolated. \nModern mu
 ltimodal AI pipelines introduce significant computational challenges becau
 se of the size and heterogeneity of medical datasets. Training multimodal 
 large models require GPU acceleration, high-memory systems, distributed st
 orage, and reproducible workflow orchestration. Bridges-2 provides an effe
 ctive environment for these workloads through its GPU-enabled architecture
 , flexible compute resources, and support for data-intensive AI applicatio
 ns. Combined with Neocortex’s AI-focused tooling and orchestration capabil
 ities, these systems support scalable experimentation and rapid prototypin
 g for biomedical machine learning research.\nThe presentation will discuss
  practical considerations for deploying multimodal AI workflows on HPC inf
 rastructure, including distributed training strategies, memory-efficient f
 ine-tuning, model parallelism, and optimization for large transformer-base
 d architectures. It will also explore challenges related to biomedical dat
 a movement, storage, and privacy-aware computing in institutional research
  environments\n\nDomain: Engineering, Life Sciences, Computational Methods
  and Applied Mathematics\n\nSession Chairs: John Anderson Garcia Henao (Un
 iversity of Bern, ARTORG Center for Biomedical Engineering Research) and C
 arlos Barrios Hernandez (Universidad Industrial de Santander, LIG/INRIA - 
 CITI Laboratory)\n\n
END:VEVENT
END:VCALENDAR