BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Machine Learning for Liquid Argon TPCs: From Task-Specific Network
 s to Foundation Models
DTSTART:20261005T164500Z
DTEND:20261005T174500Z
DTSTAMP:20261018T112100Z
UID:indico-event-1733534@indico.cern.ch
DESCRIPTION:Speakers: Matteo Vicenzi (Brookhaven National Laboratory (US))
 \n\nLiquid argon time projection chambers (LArTPCs) are the central detect
 or technology of the current and next-generation neutrino program in the U
 S. These detectors record neutrino interactions as high-resolution 2D imag
 es which are then combined into 3D pictures of each event. Over the past f
 ew years\, traditional reconstruction frameworks have progressively integr
 ated machine learning approaches for specific tasks. This has also been th
 e case for Wire-Cell\, a mature data-first reconstruction framework develo
 ped at BNL. These trends have accelerated rapidly. In this talk\, I will r
 eview some recent applications of machine learning for LArTPCs\, including
  full deep-learning reconstruction chains\, surrogate models and different
 iable simulations\, which open new approaches to calibration and inference
 . Most of these methods rely on supervised training with simulated data. F
 oundation models offer a complementary approach\, learning general-purpose
  representations directly from detector data through self-supervised train
 ing\, which can then be adapted to a wide range of downstream tasks. I wil
 l conclude by discussing the ongoing development of a Wire-Cell Foundation
  Model\, including data preparation\, model architecture\, initial trainin
 g\, and preliminary studies of its learned representations.\n\nhttps://ind
 ico.cern.ch/event/1733534/
URL:https://indico.cern.ch/event/1733534/
END:VEVENT
END:VCALENDAR