BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Cosmological parameter inference using neural networks - Tom Charn
 ock\, Institut d'Astrophysique de Paris
DTSTART:20191202T130000Z
DTEND:20191202T140000Z
UID:TALK134737@talks.cam.ac.uk
CONTACT:William Coulton
DESCRIPTION:With the information from the theoretically well-understood CM
 B almost exhausted\, we need to look for new sources to learn more about t
 he cosmological model of our universe. Whilst these new sources are plenti
 ful\, they are technically very difficult to extract information from. Neu
 ral networks with large training sets are currently providing tighter cons
 traints on cosmological parameters than ever before. However\, in their cu
 rrent form\, these neural networks are unable to give true Bayesian infere
 nce of cosmological model parameters. I will describe why this is true and
  present two methods by which the information extracting power of neural n
 etworks can be built into the necessary robust statistical framework to pe
 rform trustworthy inference\, whilst at the same time massively reducing t
 he quantity of training data required.
LOCATION:CMS\, Pav. B\, CTC Common Room (B1.19) [Potter Room]
END:VEVENT
END:VCALENDAR