Speaker
Description
One current focus point of calorimeter reconstruction algorithm developments has been the utilisation of time information as the fifth dimension of calorimeter data.
In this talk, recent work from the CALO5D team is presented, which studies timing in the highly granular CALICE-type calorimeter. While we study this calorimeter concept as implemented in the International Large Detector (ILD) concept proposed for ILC and FCC-ee, it has also been implemented as HGCAL in CMS's upgrade for the High-Luminosity phase of the LHC.
Making use of the extensive reconstruction software framework and MC data in full-simulation of ILD, novel machine-learning-based approaches to ParticleFlow are presented, with a particular focus on jet energy resolution optimisation, its composition and software compensation.
As overarching topic the impact of adding time information compared to ILD's conventional approach of Pandora ParticleFlow is discussed. Cross-detector developments in machine learning are highlighted, such as the application of GravNet, developed for the CMS HGCAL, to ILD MC data.
