Parameter inference from event ensembles and the topquark mass
Abstract
One of the key tasks of any particle collider is measurement. In practice, this is often done by fitting data to a simulation, which depends on many parameters. Sometimes, when the effects of varying different parameters are highly correlated, a large ensemble of data may be needed to resolve parameterspace degeneracies. An important example is measuring the topquark mass, where other physical and unphysical parameters in the simulation must be profiled when fitting the topquark mass parameter. We compare four different methodologies for topquark mass measurement: a classical histogram fit similar to one commonly used in experiment augmented by softdrop jet grooming; a 2D profile likelihood fit with a nuisance parameter; a machinelearning method called DCTR; and a linear regression approach, either using a leastsquares fit or with a dense linearlyactivated neural network. Despite the fact that individual events are totally uncorrelated, we find that the linear regression methods work most effectively when we input an ensemble of events sorted by mass, rather than training them on individual events. Although all methods provide robust extraction of the topquark mass parameter, the linear network does marginally best and is remarkably simple. For the top study, we conclude that the MonteCarlobased uncertainty on current extractions of the topquark mass from LHC data can be reduced significantly (by perhaps a factor of 2) using networks trained on sorted event ensembles. More generally, machine learning from ensembles for parameter estimation has broad potential for collider physics measurements.
 Publication:

Journal of High Energy Physics
 Pub Date:
 September 2021
 DOI:
 10.1007/JHEP09(2021)058
 arXiv:
 arXiv:2011.04666
 Bibcode:
 2021JHEP...09..058F
 Keywords:

 HadronHadron scattering (experiments);
 Top physics;
 High Energy Physics  Phenomenology
 EPrint:
 v1: 27 + 5 pages, 14 + 3 figures. v2: Matches version accepted to JHEP