Nukleárny seminár - Marián Ivanov (15.12.2021)
v stredu 15.12.2021 o 14:00 hod. online formou
Od: Jaroslav Staníček
Pozývame Vás na Nukleárny seminár Katedry jadrovej fyziky a biofyziky FMFI UK, Slovenskej fyzikálnej spoločnosti a Slovenskej nukleárnej spoločnosti
Prednášajúci: Dr. Marián Ivanov (CERN, GSI Darmstadt, University Heidelberg)
Názov: RootInteractive tool pre N dimenzionálnu interaktívnu analýzu: Vidieť znamená veriť
Termín: 15.12.2021, 14:00 hod., MS Teams
ALICE, one of the four big experiments at the CERN LHC, is a detector dedicated to heavy-ion physics. A high interaction rate environment causes pile-up which necessitates the use of advanced methods of data analysis. Over the recent years machine learning (ML) has come to dominate multi-dimensional data analysis. However, it is more difficult to interpret the ML models and to evaluate their uncertainties, which are offered by classical approaches.
In this seminar we would like to present new tool RootInteractive which we recently developed within ALICE. The main goal of the tool is to simplify data analysis in many dimensions, to fit (ML regression) and visualize N dimensional function and to enable easy functional composition using non-parametric and parametric functions. Using RootInteractive tool strongly simplifies multidimensional parameter optimization, we enabled interactive expert communication within ALICE working groups. The tool is aimed not only for computer experts but is used by "standard” users (master, PhD) and for the educational purposes.
We will show how RootInteractive is used in ALICE for reconstruction, calibration, MC simulations and also preliminary physics analysis studies. In more detail, we will demonstrate how we combine ML with a parametric model, in order to yield a compact representation of physics processes. Our main use case is the calibration of space charge distortions, which requires estimates of reducible and irreducible uncertainties. We will demonstrate how this and other use cases (Particle Identification, V0/Cascade reconstruction, MC/data remapping) are solved with our approach and will describe the features of the software we developed.