BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:IAP - High-Energy Universe Group Seminar (Igor Vaiman\,  UHECR and
  IFT)
DTSTART:20260507T090000Z
DTEND:20260507T103000Z
DTSTAMP:20260611T185600Z
UID:indico-event-5600@indico.kit.edu
DESCRIPTION:WS 2025/26\nIgor Vaiman\nGSSI\, Italy\n" Reconstruction of UHE
 CR source spectra with Information Field Theory"\nabstract: The Pierre Aug
 er Observatory measures key characteristics of ultra-high-energy cosmic ra
 ys (UHECRs): energy\, depth of the shower maximum (sensitive to mass)\, an
 d arrival direction. The former two are known to constrain the properties 
 of cosmic-ray sources in a simplified case of diffuse\, spherically symmet
 ric\, identical sources. However\, the results are often challenging for t
 heoretical interpretation (e.g.\, the very hard spectral index down to -1.
 5) and sensitive to assumptions of the model (e.g.\, cutoff shape and sour
 ce evolution with distance). We show that the parameter space of standard 
 spectral models is highly degenerate and doesn’t offer a clear insight i
 nto the underlying physics. This motivates a nonparametric approach\, for 
 which we use methods and ideas from Information Field Theory. We model the
  injection energy spectrum as a Gaussian random field\, assuming a certain
  apriori correlation structure to ensure physical interpretability of the 
 reconstruction. We benchmark the method using both synthetic and real data
  in well-established scenarios: the dip model for the proton-only scenario
 \, and the full combined fit of UHECR spectrum and composition. The recons
 truction performs well on synthetic data\, reproduces main results from th
 e literature\, and manages to find better-fitting source spectra for most 
 cases. We also discuss further extension of the analysis to allow the sour
 ce spectrum to vary with distance.\nchair: Michael Unger\n \n \n \n \n
 \nhttps://indico.kit.edu/event/5600/
LOCATION:206 (KIT CN Bld 425)
URL:https://indico.kit.edu/event/5600/
END:VEVENT
END:VCALENDAR