BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Real time AI for particle physics measurements
DTSTART:20260716T140000Z
DTEND:20260716T150000Z
DTSTAMP:20260713T012400Z
UID:indico-event-898@indico.physik.uni-siegen.de
DESCRIPTION:Speakers: Torben Ferber\n\nWhat collider experiments can ultim
 ately measure is often constrained by what can be reconstructed and select
 ed in real time\, within a fixed latency of a few microseconds. \nAt Bell
 e II this is especially true for dark-sector searches\, low-multiplicity t
 au decays\, and final states with few tracks or photons\, where background
 s overwhelm the signal and the trigger must decide what to keep before any
  data is stored.\nThe talk centers on three examples at Belle II. \nThe f
 irst is multi-modal GNN-based track finding\, originally developed to reco
 ver sensitivity to long-lived particles\, whose displaced decays break the
  prompt-track assumptions of conventional pattern recognition. \nThe seco
 nd is real-time\, GNN-based clustering in the electromagnetic calorimeter 
 as it already operates today on FPGAs\, within the trigger latency budget 
 of 1.05 microseconds. \nFinally I will look ahead to a future calorimeter
  trigger upgrade with substantially more inputs\, where the same class of 
 architecture is a candidate for a harder clustering problem. \nDeploying 
 any of this remains the bottleneck: I will describe the chain from trained
  model to inference on heterogeneous platforms such as AMD Versal\, and th
 e worldwide efforts\, including the ErUM-Data DEEP project\, to simplify a
  process that is still genuinely difficult.\n\nhttps://indico.physik.uni-s
 iegen.de/event/898/
URL:https://indico.physik.uni-siegen.de/event/898/
END:VEVENT
END:VCALENDAR