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 26.5
DTSTART;TZID=Europe/Stockholm:20240604T120000
DTEND;TZID=Europe/Stockholm:20240604T123000
UID:submissions.pasc-conference.org_PASC24_sess162_msa116@linklings.com
SUMMARY:Privacy-Preserving Federated Learning-as-a-Service: Building Trust
 worthy AI Models and Biomedical Insights
DESCRIPTION:Minisymposium\n\nRavi Madduri (Argonne National Laboratory, Un
 iversity of Chicago)\n\nFederated learning (FL) is a collaborative learnin
 g approach where multiple data owners train a model together under the orc
 hestration of a central server by sharing the model trained on their local
  datasets instead of sharing the data directly. FL enables creation of mor
 e robust models without the exposure of local datasets. However, FL by its
 elf, does not guarantee the privacy of data, because the information extra
 cted from the communication of FL algorithms can be accumulated and utiliz
 ed to infer the private local data used for training. We developed Advance
 d Privacy Preserving Federated Learning framework (APPFL), with advances i
 n differential privacy, to enable Privacy-Preserving Federated Learning (P
 PFL). We enabled PPFL through scaled, distributed training on supercomputi
 ng resources across multiple institutions to help create robust, trust-wor
 thy AI models in biomedicine and smart grid applications. Setting up a sec
 ure high-performance computing FL experiment requires capabilities that ma
 y not be available for all. To lower the barrier for leveraging PPFL, we c
 reated the Advanced Privacy-Preserving Federated Learning as a service (AP
 PFLx). APPFLx enables cross-silo PPFL with easy to use web interface for m
 anaging, deploying, analyzing, and visualizing PPFL experiments. In this t
 alk, we will describe APPFLx and its adoption to biomedical use cases.\n\n
 Domain: Applied Social Sciences and Humanities, Life Sciences, Computation
 al Methods and Applied Mathematics\n\nSession Chair: Destinee Morrow (Lawr
 ence Berkeley National Laboratory)
END:VEVENT
END:VCALENDAR