BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Theory uncertainties and missing higher orders at LHC
DTSTART:20260420T120000Z
DTEND:20260420T170000Z
DTSTAMP:20260503T013300Z
UID:indico-event-49467@agenda.infn.it
DESCRIPTION:Speakers: Mauro Valli (Istituto Nazionale di Fisica Nucleare)\
 , Valerio Ippolito (Istituto Nazionale di Fisica Nucleare)\, Francesco Pan
 dolfi (INFN Rome)\n\nThe topic of this afternoon is the quantification of 
 theory uncertainty due to the missing higher orders in perturbative predic
 tions. The canonical approach\, based on the variation of unphysical scale
 s\, suffers from several limitations and may lead to unreliable estimates 
 of these uncertainties\, particularly in their correlations. Moreover\, it
  lacks a clear probabilistic foundation.\nThe talks will present new appro
 aches aimed at overcoming the limitations of scale variation. After a gene
 ral introduction to probabilistic methods in high-energy physics and to th
 e comparison between data and theory\, alternative and powerful techniques
  to quantify theoretical uncertainties based on a Bayesian framework will 
 be discussed.\n\nA novel approach based on theory nuisance parameters (TNP
 s) will then be introduced. This method consistently accounts for theoreti
 cal correlations and enables the data itself to constrain\, and potentiall
 y reduce\, these uncertainties. To illustrate the potential of this framew
 ork\, an ongoing project on the extraction of $\\alpha_s(m_Z)$ and its unc
 ertainty from the Z $q_T$ spectrum through fits to pseudodata will be pres
 ented. In addition\, the recent CMS $m_W$ determination\, which successful
 ly employed TNPs as part of its uncertainty modelling\, will be briefly di
 scussed.\nFinally\, a modified version of the TNP approach that benefits f
 rom the Bayesian methods previously discussed will be introduced.\n\n\nhtt
 ps://agenda.infn.it/event/49467/
LOCATION:Aula Conversi (Dip. di Fisica - Edificio G. Marconi)
URL:https://agenda.infn.it/event/49467/
END:VEVENT
END:VCALENDAR