BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Decoding the cosmos [Confirmed]
DTSTART:20250423T180000Z
DTEND:20250423T193000Z
DTSTAMP:20260617T133700Z
UID:indico-event-1133@events.perimeterinstitute.ca
DESCRIPTION:Speakers: Hiranya Peiris (Cambridge\, IoA)\n\nCosmology is und
 ergoing a data revolution. Surveys such as the imminent Legacy Survey of S
 pace and Time (LSST) to be conducted by the Vera C. Rubin Observatory will
  deliver huge galaxy catalogues that provide critical tools for understand
 ing the nature of dark matter and dark energy. However\, in order to obtai
 n accurate cosmological constraints from these enormous datasets\, we need
  reliable ways of estimating galaxy properties using only photometry. I wi
 ll present pop-cosmos: a forward modelling framework for photometric galax
 y survey data\, where galaxies are modelled as draws from a population pri
 or distribution over redshift\, mass\, dust properties\, metallicity\, and
  star formation history. These properties are mapped to photometry using a
 n emulator for stellar population synthesis\, followed by the application 
 of a learned model for a survey's noise properties. Application of selecti
 on cuts enables the generation of mock galaxy catalogues. This enables us 
 to use simulation-based inference to solve the inverse problem of calibrat
 ing the population-level prior on a deep multiwavelength catalogue\, COSMO
 S2020. We use a diffusion model as a flexible population-level prior\, and
  optimise its parameters by minimising the Wasserstein distance between fo
 rward-simulated photometry and the real COSMOS2020 survey data. The result
 ing model can then be used to derive accurate redshift distributions for u
 pcoming photometric surveys\, to facilitate weak lensing and clustering sc
 ience. I will show applications of this framework\, demonstrating how we a
 re able to extract redshift distributions\, and make inferences about gala
 xy evolution. I will also discuss the use of pop-cosmos as a prior for per
 forming inference on individual galaxies in a highly scaleable manner\, as
  well as ongoing work to analyse data from the Kilo-Degree Survey (KiDS) i
 n preparation for LSST.\n\nhttps://events.perimeterinstitute.ca/event/1133
 /
LOCATION:PI/2-292 - Time Room (Perimeter Institute for Theoretical Physics
 )
URL:https://events.perimeterinstitute.ca/event/1133/
END:VEVENT
END:VCALENDAR