Conveners
PD4: Software & Detector Performance
- Manqi Ruan (Chinese Academy of Sciences (CN))
- Adrian Irles (IFIC CSIC/UV)
- Andre Sailer (CERN)
- Daniel Jeans (KEK)
- Jan Strube (PNNL)
- Frank-Dieter Gaede (Deutsches Elektronen-Synchrotron (DE))
- Graham Wilson (The University of Kansas (US))
PD4: Software & Detector Performance
- Graham Wilson (The University of Kansas (US))
- Adrian Irles (IFIC CSIC/UV)
- Andre Sailer (CERN)
- Daniel Jeans (KEK)
- Frank-Dieter Gaede (Deutsches Elektronen-Synchrotron (DE))
- Jan Strube (PNNL)
- Manqi Ruan (Chinese Academy of Sciences (CN))
A new LCIO-based data format called mini-DST has been developed, which combines PFO- and event-level information, including the output of the most important high-level reconstruction algorithms.
Originally triggered by Snowmass 2021 studies, the mini-DST is useful for beginners as the starting point of analysis.
In this talk, we discuss the basics and contents of the mini-DST, how to use it...
The Event Data Model (EDM) is at the core of HEP experiments software frameworks. It defines the language in which physicists are able to express their ideas and also how the different software components communicate with each other. The Key4HEP project aims to develop a common software stack for future collider projects. One of the main components of Key4HEP is a common EDM in the form of...
Since 2020, CLIC and ILC take part in the Key4hep collaboration, which strives to create a common software for HEP collider design studies. Key4hep represents a flexible, multi-layered model of collaboration, where different common components like documentation, build system, data modeling, persistency and framework components are adopted as needed. This talks gives a bird's-eye view of...
To ensure the backward compatibility between iLCSoft and Key4hep and to ease the validation of the iLCSoft processors, k4MarlinWrapper provides the necessary tools to run Marlin processors using the Gaudi framework, allowing for a smooth transition from current battle-tested particle reconstruction frameworks, to a common framework for future experiments like CLIC or FCC. It creates a wrapper...
The iLCDirac grid interface has been successfully used by Linear Collider community for many years. It has made it possible to isolate the users from the ever changing distributed environments by offering a consistent interface throughout the years. In this contribution we detail the current status and latest developments as well as the plans for keeping iLCDirac up-to-date with the latest...
We have produced new high statistics 250 GeV common MC samples
for the ILD physics study using the latest generator, simulation, and reconstruction packages.
Aiming for the requested statistics of MC samples for physics study,
we utilized ILCDIRAC distributed computing environment for mass production.
In this talk, we will report the estimated resource requirements
and the current...
We are developing kinematic fitter which can deal with arbitrary resolution functions. Kinematic fitting is the constrained optimization method which uses distributions of fit parameters and kinematic relations among the parameters. In order to treat non-Gaussian distributions, for example b-jet energy distribution, our kinematic fitter is implemented based on the log-likelihood method.
In...
In this talk, we show the recent results of our R&D works on the Machine Learning Application to the Collider Experiments.
In RCNP, Osaka University, in Japan, we form a group which consists with about 20 researchers on both information science and collider physics (experiment and theory) to work on the R&D of machine learning application to the collider experiments, as a research project in...
Jet clustering is one of the main key to obtain better physics results because
reducing mis-clustring leads to improve the mass resolution of the resonances especially in multi-jet situation.
Present jet clustering is far from a good tool for reconstructing jets. We need to tackle the problem
and should explore the possibility of constructing better jet clustering algorithm.
Recently,...
We developed a novel algorithm of vertex finding for future lepton colliders such as the International Linear Collider. We deploy two networks; one is simple fully-connected layers to look for vertex seeds from track pairs, and the other is a customized Recurrent Neural Network with an attention mechanism and an encoder-decoder structure to associate tracks to the vertex seeds. The performance...
Tau lepton physic plays an important role in the research programme at future e+e- experiments. To fully exploit the physics potential of machine and experiments, and for a cost-effective detector design, it is important to to implement from start advanced Machine Learning methods in the development of the detector. With this respect we report here on an ongoing study on τ-identification...
Accurate simulation of physical processes is
crucial for the success of modern particle physics.
However, simulating the development and interaction of particle showers with calorimeter detectors is a time consuming process and drives the computing needs of large experiments at the LHC and future colliders. Recently, generative machine
learning models based on deep neural networks have...
