BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Reconstructing the Type Ia Supernova Absolute Magnitude with Two-P
 robe Physics-Informed Neural Networks
DTSTART;VALUE=DATE-TIME:20260916T120000Z
DTEND;VALUE=DATE-TIME:20260916T122500Z
DTSTAMP;VALUE=DATE-TIME:20261023T210728Z
UID:indico-contribution-31296@indico.ific.uv.es
DESCRIPTION:Speakers: Denitsa Staicova (Institute for Nuclear Research and
  Nuclear Energy)\nWe apply two variants of Physics-Informed Neural Network
 s (PINNs) to reconstruct the Type Ia supernova absolute magnitude M_B(z) f
 rom joint BAO and supernova data under four cosmological models (\\LambdaC
 DM\, CPL\, GEDE\, \\Lambda_sCDM) and two DESI DR2 fiducial sets. A heteros
 cedastic single-network method tested across four constraint configuration
 s establishes that the Etherington distance duality relation is more funda
 mental constraint than cosmological model priors\, with DDR violations of 
 30--52 mmag under physical constraints versus 85--2330 mmag without. Under
  full constraints all models recover M_B \\approx -19.3 mag with biases be
 low 0.05 mag. A Fisher information-weighted two-network variant trains ind
 ependent networks on BAO and SN data\, providing clean probe separation an
 d finding no significant M_B evolution in z \\in [0.3\, 1.5]. The heterosc
 edastic method identifies a persistent 2-3\\sigma residual at z \\sim 0.4-
 0.5 that is consistent across all four models and both fiducials\; the Fis
 her method finds no significant pointwise deviation in z\\in[0.3\,1.5] but
  shows a systematic separation of redshift-binned M_B distributions consis
 tent with the same underlying tension. While the origin of this feature re
 mains ambiguous\, its model-independence and cross-method consistency warr
 ant further investigation with forthcoming data.\n\nhttps://indico.ific.uv
 .es/event/8160/contributions/31296/
LOCATION:
URL:https://indico.ific.uv.es/event/8160/contributions/31296/
END:VEVENT
END:VCALENDAR