arXiv subjects

Federico Colecchia

Publications and source records attributed to Federico Colecchia.

7 recordsLinked to original sources

arXiv2017

On the frequency distribution of neutral particles from low-energy strong interactions

The rejection of the contamination, or background, from low-energy strong interactions at hadron collider experiments is a topic that has received significant attention in the field of particle physics. This article builds on a particle-level view of collision events, in line with recently-proposed subtraction methods. While conventional techniques in the field usually concentrate on probability distributions, our study is, to our knowledge, the first attempt at estimating the frequency distribution of background particles across the kinematic space inside individual collision events. In fact, while the probability distribution can generally be estimated given a model of low-energy strong interactions, the corresponding frequency distribution inside a single event typically deviates from the average and cannot be predicted a priori. We present preliminary results in this direction, and establish a connection between our technique and the particle weighting methods that have been the subject of recent investigation at the Large Hadron Collider.

hep-ph↗

arXiv2015

Data-driven estimation of neutral pileup particle multiplicity in high-luminosity hadron collider environments

The upcoming operation regimes of the Large Hadron Collider are going to place stronger requirements on the rejection of particles originating from pileup, i.e. from interactions between other protons. For this reason, particle weighting techniques have recently been proposed in order to subtract pileup at the level of individual particles. We describe a choice of weights that, unlike others that rely on particle proximity, exploits the particle-level kinematic signatures of the high-energy scattering and of the pileup interactions. We illustrate the use of the weights to estimate the number density of neutral pileup particles inside individual events, and we elaborate on the complementarity between ours and other methods. We conclude by suggesting the idea of combining different sets of weights with a view to exploiting different features of the underlying processes for improved pileup subtraction at higher luminosity.

hep-ph↗

arXiv2014

A sampling algorithm to estimate the effect of fluctuations in particle physics data

Background properties in experimental particle physics are typically estimated using large data sets. However, different events can exhibit different features because of the quantum mechanical nature of the underlying physics processes. While signal and background fractions in a given data set can be evaluated using a maximum likelihood estimator, the shapes of the corresponding distributions are traditionally obtained using high-statistics control samples, which normally neglects the effect of fluctuations. On the other hand, if it was possible to subtract background using templates that take fluctuations into account, this would be expected to improve the resolution of the observables of interest, and to reduce systematics depending on the analysis. This study is an initial step in this direction. We propose a novel algorithm inspired by the Gibbs sampler that makes it possible to estimate the shapes of signal and background probability density functions from a given collection of particles, using control sample templates as initial conditions and refining them to take into account the effect of fluctuations. Results on Monte Carlo data are presented, and the prospects for future development are discussed.

physics.data-an↗

arXiv2014

Toward particle-level filtering of individual collision events at the Large Hadron Collider and beyond

Low-energy strong interactions are a major source of background at hadron colliders, and methods of subtracting the associated energy flow are well established in the field. Traditional approaches treat the contamination as diffuse, and estimate background energy levels either by averaging over large data sets or by restricting to given kinematic regions inside individual collision events. On the other hand, more recent techniques take into account the discrete nature of background, most notably by exploiting the presence of substructure inside hard jets, i.e. inside collections of particles originating from scattered hard quarks and gluons. However, none of the existing methods subtract background at the level of individual particles inside events. We illustrate the use of an algorithm that can enable particle-by-particle background discrimination at the Large Hadron Collider, and we envisage this as the basis for a novel event filtering procedure upstream of the official jet reconstruction pipelines. Our hope is that this new technique will improve physics analysis when used in combination with state-of-the-art algorithms in high-luminosity hadron collider environments.

physics.data-an↗

arXiv2014

Toward the estimation of background fluctuations under newly-observed signals in particle physics

When the number of events associated with a signal process is estimated in particle physics, it is common practice to extrapolate background distributions from control regions to a predefined signal window. This allows accurate estimation of the expected, or average, number of background events under the signal. However, in general, the actual number of background events can deviate from the average due to fluctuations in the data. Such a difference can be sizable when compared to the number of signal events in the early stages of data analysis following the observation of a new particle, as well as in the analysis of rare decay channels. We report on the development of a data-driven technique that aims to estimate the actual, as opposed to the expected, number of background events in a predefined signal window. We discuss results on toy Monte Carlo data and provide a preliminary estimate of systematic uncertainty.

physics.data-an↗

arXiv2014

Particle-level kinematic fingerprints and the multiplicity of neutral particles from low-energy strong interactions

The contamination, or background, from uninteresting low-energy strong interactions is a major issue for data analysis at the Large Hadron Collider. In the light of the challenges associated with the upcoming higher-luminosity scenarios, methods of assigning weights to individual particles have recently started to be used with a view to rescaling the particle four-momentum vectors. We propose a different approach whereby the weights are instead employed to reshape the particle-level kinematic distributions in the data. We use this method to estimate the number of neutral particles originating from low-energy strong interactions in different kinematic regions inside individual collision events. Given the parallel nature of this technique, we anticipate the possibility of using it as part of particle-by-particle event filtering procedures at the reconstruction level at future high-luminosity hadron collider experiments.

hep-ph↗

arXiv2013

A population-based approach to background discrimination in particle physics

Background properties in experimental particle physics are typically estimated using control samples corresponding to large numbers of events. This can provide precise knowledge of average background distributions, but typically does not consider the effect of fluctuations in a data set of interest. A novel approach based on mixture model decomposition is presented as a way to estimate the effect of fluctuations on the shapes of probability distributions in a given data set, with a view to improving on the knowledge of background distributions obtained from control samples. Events are treated as heterogeneous populations comprising particles originating from different processes, and individual particles are mapped to a process of interest on a probabilistic basis. The proposed approach makes it possible to extract from the data information about the effect of fluctuations that would otherwise be lost using traditional methods based on high-statistics control samples. A feasibility study on Monte Carlo is presented, together with a comparison with existing techniques. Finally, the prospects for the development of tools for intensive offline analysis of individual events at the Large Hadron Collider are discussed.

physics.data-an↗