BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Neural ratio estimation: the future of supernova cosmology? - Matt
 hew Grayling (University of Cambridge)
DTSTART:20240424T090000Z
DTEND:20240424T100000Z
UID:TALK216250@talks.cam.ac.uk
CONTACT:David Yallup
DESCRIPTION:Simulation-based inference (SBI) has the potential to revoluti
 onise how we do supernova cosmology and let us incorporate arbitrarily com
 plex effects within a Bayesian model. I will present recent work which sou
 ght to validate neural ratio estimation (NRE) by comparing NRE-derived pos
 teriors on supernova properties to those obtained with a likelihood-based 
 MCMC approach for the same data\, and then discuss how NRE and SBI in gene
 ral provide a pathway towards a model extending all the way from type Ia s
 upernova light curves to cosmological parameters as part of a single analy
 sis.
LOCATION:Martin Ryle Seminar Room\, KICC
END:VEVENT
END:VCALENDAR