BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//juliacon-2026//talk//ZLZ8FJ
BEGIN:VTIMEZONE
TZID:Europe/Berlin
BEGIN:DAYLIGHT
DTSTART:20250813T000000
TZNAME:CEST
TZOFFSETFROM:+0200
TZOFFSETTO:+0200
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251026T030000
RDATE:20261025T030000
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20260329T030000
RDATE:20270328T030000
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Algorithmic differentiation and error control with DFTK - Bruno Pl
 oumhans
DTSTART;TZID=Europe/Berlin:20260813T161500
DTEND;TZID=Europe/Berlin:20260813T163000
DTSTAMP:20260812T152510Z
UID:pretalx-juliacon-2026-ZLZ8FJ@pretalx.com
DESCRIPTION:The Density-Functional ToolKit (DFTK) is a Julia package provi
 ding routines to compute the electronic structure of a bulk material and r
 elated properties\, using plane-wave density functional theory (DFT). Many
  material properties of interest can be expressed as derivatives of simula
 tion outputs wrt. input parameters\, and typically only specific combinati
 ons are implemented by DFT codes\, as a result of great programming effort
  to hand-implement all the required derivative terms. In DFTK however\, de
 rivatives of **any** output quantity wrt. **any** input parameter can be c
 omputed\, using algorithmic differentiation (AD) combined with density-fun
 ctional perturbation theory (DFPT). This results in a general AD-DFPT fram
 ework [1] that can only be used to compute both standard and novel derivat
 ives\, with promising applications including gradient-based optimization a
 nd error propagation.\n\nIn the first part of this talk\, I will discuss t
 he key ideas behind this implementation\, showing how we offload tedious d
 erivative computations to the AD framework\, while keeping the numerics un
 der control thanks to the underlying DFPT solver. The overall strategy is 
 quite general\, and should be applicable in other fields as well. In the s
 econd part of this talk\, I will present new research directions enabled b
 y AD-DFPT. In particular\, I will focus on the propagation of model parame
 ter uncertainty \nand estimated numerical errors all the way to predicted 
 physical quantities.\n\n[1]: N. F. Schmitz\, B. Ploumhans\, M. F. Herbst. 
 _npj Comput Mater_ **12**\, 6 (2026). https://doi.org/10.1038/s41524-025-0
 1880-3
LOCATION:Muschel — N2
URL:https://pretalx.com/juliacon-2026/talk/ZLZ8FJ/
END:VEVENT
END:VCALENDAR