BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//juliacon-2026//speaker//KGWJBD
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-juliacon-2026-TGC3ZM@pretalx.com
DTSTART;TZID=CET:20260814T121500
DTEND;TZID=CET:20260814T123000
DESCRIPTION:Electron Probe Microanalysis (EPMA) is an imaging technique for
  the quantitative analysis of solid material samples relying on measuremen
 ts of characteristic X-ray emission induced by electron irradiation.\nThe 
 determination of the material constitutes an inverse problem\, hence an ef
 ficient reconstruction requires differentiability of the forward model.\nT
 he mathematical model employed in EPMA is governed by a linear transport e
 quation that for heterogeneous materials is commonly approximated using Mo
 nte Carlo simulation\, where statistical noise complicates the computation
  of gradients.\nFor reconstruction\, there exist surrogate models that are
  well tested in practice\, but are very restrictive in the parametrization
  of the material\, allowing only homogeneous or depth-layered materials\, 
 which ultimately limits the spatial resolution of quantitative analysis in
  EPMA.\n\nIn this short talk\, we present an implementation of a determini
 stic\, heterogeneous\, and differentiable model for EPMA in Julia.\nRecons
 truction can then be implemented as a gradient-based optimization using th
 e model as a PDE constraint.\nCompatibility with algorithmic differentiati
 on allows us to tailor the material parametrization to a set of quantities
  of interest\, depending on the requirements of a specific sample.\nRecons
 truction results using realistic as well as synthetic measurements demonst
 rate potential for further development.\n\nAdditionally\, we briefly discu
 ss a structural similarity of the forward model in EPMA to the structure i
 n which adjoint methods can be effectively applied for gradient computatio
 n. It allows us to "apply adjoints twice" leading also to a more efficient
  computation of the forward problem\, ultimately accelerating reconstructi
 on approaches.
DTSTAMP:20260710T083552Z
LOCATION:Alte Mensa — Atrium Maximum
SUMMARY:How implementing a differentiable model for Electron Microscopy (EP
 MA) accelerated the forward simulation - Tamme Claus
URL:https://pretalx.com/juliacon-2026/talk/TGC3ZM/
END:VEVENT
END:VCALENDAR