2024-03-28T21:37:39Z
https://zenodo.org/oai2d
oai:zenodo.org:4147617
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Schwinzerl, Martin
2020-07-17
<p>Python is increasingly becoming the language of choice for a wide range of applications within the scientific community. For performance critical and high-performance computing (HPC) related tasks, however, it is still beneficial to develop critical code paths in a statically typed language like C or C++ and to then provide bindings to Python. In this talk, we present different options (e.g. ctypes, pybind11, cppyy) to provide these bindings, using both a simple demo library and SixTrackLib, an existing real-world high-performance C/C++ library as examples.</p>
<p>SixTrackLib in particular features multiple computing back-ends and a complex set of dependencies and configuration options which have to remain accessible to the users of the Python bindings. We compare the ease of use, support for modularity, and resulting performance across the different binding strategies. Moreover, we consider the interaction of the binding generation process with a build-system like CMake. We devote special attention to strategies and practices ensuring the consistent application of compiler options and sketch strategies for seamlessly performing library installations.</p>
https://doi.org/10.5281/zenodo.4147617
oai:zenodo.org:4147617
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147616
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Providing Python Bindings For Complex and Feature-Rich C and C++ Libraries
info:eu-repo/semantics/lecture
oai:zenodo.org:4147308
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Straub, David
2020-07-14
<p>Keynote presentation.</p>
https://doi.org/10.5281/zenodo.4147308
oai:zenodo.org:4147308
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147307
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Python & HEP: a perfect match, in theory
info:eu-repo/semantics/lecture
oai:zenodo.org:4147548
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Bendavid, Josh
2020-07-16
<p>With increasing integrated luminosity at the LHC, highly differential and extremely precise measurements of Standard Model processes become possible. This imposes stringent requirements for accurate modelling of systematic uncertainties of both experimental and theoretical origin. One possible method for the unfolding of detector response is the use of maximum likelihood fits, with systematic uncertainties represented by nuisance parameters. Precise Standard Model measurements in Run 2 already pose a challenge with respect to the speed and stability of Standard Tools. A Tensorflow-based implementation for binned maximum likelihood fits has been used for the unfolding and statistical interpretation of high precision differential cross section measurements of W boson production recently released by CMS, involving maximum likelihood fits including hundreds of millions of events, with hundreds of measured cross sections and over a thousand nuisance parameters. This has led to huge benefits in terms of speed, stability, and flexibility for the sophistication of the measurement and systematic uncertainties, as well as statistical interpretation.</p>
https://doi.org/10.5281/zenodo.4147548
oai:zenodo.org:4147548
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147547
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Tensorflow-based Maximum Likelihood fits for High Precision Standard Model Measurements at CMS
info:eu-repo/semantics/lecture
oai:zenodo.org:4147607
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Oeftiger, Adrian
2020-07-17
<p>I have a high-performance number crunching tool with cool physics which simulates long-term on a GPU -- how can I extend the inner loop by further cool physics, injected from the outside?<br>
In python this should be easy, right? But wait... we are sitting on device memory?</p>
<p>In this talk we explore how to tightly couple two libraries for high-performance computation of long-term beam dynamics, SixTrackLib and PyHEADTAIL. How can we design the \emph{interface} between both libraries in terms of (1) remaining on the python level, (2) avoid losing performance due to device-to-host-to-device copies, and (3) keeping both libraries as stand-alone packages?</p>
<p>The interface can be surprisingly simple, yet fully fledged... Let's go!</p>
https://doi.org/10.5281/zenodo.4147607
oai:zenodo.org:4147607
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147606
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Integrating GPU libraries for fun and profit
info:eu-repo/semantics/lecture
oai:zenodo.org:4067878
2020-10-08T00:26:55Z
openaire
user-pyhep2020
Schreiner, Henry Fredrick
2020-07-16
<p>Hands-on tutorial with notebooks.</p>
https://doi.org/10.5281/zenodo.4067878
oai:zenodo.org:4067878
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4067877
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
High-performance Python
info:eu-repo/semantics/lecture
oai:zenodo.org:4067099
2020-10-08T10:37:14Z
software
user-pyhep2020
Lukas Heinrich
2020-10-05
<p>Tutorial on the automatic differentiation.</p>
https://doi.org/10.5281/zenodo.4067099
oai:zenodo.org:4067099
Zenodo
https://github.com/lukasheinrich/pyhep2020-autodiff-tutorial/tree/0.0.2
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4067098
info:eu-repo/semantics/openAccess
Other (Open)
lukasheinrich/pyhep2020-autodiff-tutorial 0.0.2
info:eu-repo/semantics/other
oai:zenodo.org:3961670
2020-07-28T00:59:24Z
openaire
user-pyhep2020
Smith, Nick
2020-07-13
<p>Notebook presentation describing the NanoEvents object developed for CMS exploiting the awkward-array package.</p>
https://doi.org/10.5281/zenodo.3961670
oai:zenodo.org:3961670
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961669
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
The NanoEvents object
info:eu-repo/semantics/lecture
oai:zenodo.org:3961637
2020-07-28T00:59:24Z
openaire
user-pyhep2020
Burton, Charles
2020-07-17
<p>Exploitation of Python for data acquisition work.</p>
https://doi.org/10.5281/zenodo.3961637
oai:zenodo.org:3961637
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961636
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Integrated Data Acquisition in Python
info:eu-repo/semantics/lecture
oai:zenodo.org:4147540
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Eschle, Jonas
Marinangeli, Matthieu
2020-07-16
<p>zfit is a model fitting library based on top of TensorFlow and built for customization. It can build models, load data, create and optimize losses. hepstats is a package for statistical inference and is build on top of the zfit interface, and can therefore use models and losses built in zfit directly.</p>
<p>In this tutorial, we propose to split the tutorial into two parts (switching speaker in-between): we first give an introduction (~30 mins) to zfit ranging from simple mass fits to more complicated examples including custom built PDFs and simultaneous fits. The second part (~15 mins) consists of an introduction to hepstats using the models and losses built before in zfit for statistical inference including limit setting and confidence intervals.</p>
<p>The tutorial is targeted towards beginners regarding the experience with zfit or hepstats.</p>
https://doi.org/10.5281/zenodo.4147540
oai:zenodo.org:4147540
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147539
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Model building and statistical inference with zfit and hepstats
info:eu-repo/semantics/lecture
oai:zenodo.org:4147528
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Eschle, Jonas
2020-07-16
<p>zfit is a model fitting library completely implemented in Python and based on the Deep Learning framework TensorFlow. With the recent release of TensorFlow 2.0, the structure of the TensorFlow library, as well as zfit, fundamentally changed; what was before a head-twisting exotic graph building library became a numpy-like, JIT compilable computational backend. It works with Numpy compatible arrays and offers an alternative to the Numpy + numba combination offering possibly more features such as GPU support and gradients. This new approach together with other developments make zfit a versatile library that fully supports now functions which use Python dynamics as well as being compatible with other, Numpy array using packages.<br>
In this tutorial, we will talk 10' with classical slides introducing TensorFlow 2.0 as a computational, Numpy-like library and the impact of this on zfit. The other 10' mins will be spent in a notebook showing some functions with TensorFlow 2.0 code as a direct alternative to Numpy + numba (and others).</p>
https://doi.org/10.5281/zenodo.4147528
oai:zenodo.org:4147528
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147527
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
zfit - TensorFlow 2.0: dynamic and compiled HPC
info:eu-repo/semantics/lecture
oai:zenodo.org:4147347
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Rieger, Marcel
2020-07-14
<p>In particle physics, workflow management systems are primarily used as tailored solutions in dedicated areas such as Monte Carlo production. However, physicists performing data analyses are usually required to steer their individual workflows manually which is time-consuming and often leads to undocumented relations between particular workloads.<br>
We present the luigi analysis workflow (law) Python package which is based on the open-source pipelining tool luigi, originally developed by Spotify. It establishes a generic design pattern for analyses of arbitrary scale and complexity, and shifts the focus from executing to defining the analysis logic. Law provides the building blocks to seamlessly integrate with interchangeable remote resources without, however, limiting itself to a specific choice of infrastructure. In particular, it introduces the paradigm of complete separation between analysis algorithms on the one hand, and run locations, storage locations, and software environments on the other hand.<br>
To cope with the sophisticated demands of end-to-end HEP analyses, law supports job execution on WLCG infrastructure (ARC, gLite) as well as on local computing clusters (HTCondor, Slurm, LSF), remote file access via most common protocols through the Grid File Access Library (GFAL2), and an environment sandboxing mechanism with support for Docker and Singularity containers. Moreover, the novel approach ultimately aims for analysis preservation out-of-the-box.<br>
Law is developed open-source and entirely experiment independent. It is successfully employed in ttH cross section measurements and searches for di-Higgs boson production with the CMS experiment.</p>
https://doi.org/10.5281/zenodo.4147347
oai:zenodo.org:4147347
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147346
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Design Pattern for Analysis Automation on Interchangeable, Distributed Resources using Luigi Analysis Workflows
info:eu-repo/semantics/lecture
oai:zenodo.org:4147366
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Bonanomi, Matteo
2020-07-15
<p>In 2027 CERN is expected to start the High-Luminosity LHC (HL-LHC) phase. HL-LHC will integrate 10 times the current luminosity, leading to a high pile-up rate and unprecedented radiation levels. In order to cope with such a harsh environment and maintain the current physics performance, a major upgrade of the LHC detectors is required. As part of the HL-LHC detector upgrade programme, the CMS experiment is developing a High Granularity Calorimeter (HGCAL) to replace the existing endcap calorimeters.</p>
<p>Beam tests play a fundamental role in the validation of the detector design and in the study of its physics performance. In a typical beam test environment, it is important to have quick access to data, in order to perform explorative analysis, data visualization for the main physical distributions and to run data quality monitoring. During the offline analysis it is very often necessary to run comparisons between data and simulations, to identify problematic features in data, to develop preselection and cleaning cuts, and to reconstruct different observables distributions to produce the final results by means of statistical analysis.</p>
<p>Jupyter Notebooks are being used more and more by physicists to face this kind of task, as they provide an interactive interface where code can be executed, documented, and the outputs can be directly produced, analyzed and visualised. The data can be access remotely without the need of downloading to a local machine.</p>
<p>After a brief introduction about the HGCAL, the talk will focus on the major benefits that come from the use of interactive notebooks during a test beam campaign and in the subsequent phase of data analysis.</p>
<p>Slides from the presentation.</p>
https://doi.org/10.5281/zenodo.4147366
oai:zenodo.org:4147366
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147365
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
High Granularity Calorimeter (HGCAL) test beam analysis using Jupyter notebooks
info:eu-repo/semantics/lecture
oai:zenodo.org:4147555
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Waltenberger, Wolfgang
2020-07-16
<p>SModelS is an automatic, public python tool for interpreting simplified-model results from searches for new physics at the LHC. It is based on a general procedure to decompose Beyond the Standard Model (BSM) collider signatures presenting a Z2 symmetry into Simplified Model topologies. Our method provides a way to cast BSM predictions for the LHC in a model independent framework, which can be directly confronted with the relevant experimental constraints in an automated fashion. Our database contains simplified models results of about 100 CMS and ATLAS publications.</p>
<p>In this notebook talk, we wish demonstrate typical usage patterns of SModelS. We aim to show how a model is input, the list of analyses results that apply is obtained, and how likelihoods are computed. In the last part of the talk we shall quickly present new features of the upcoming v2.0 release of SModelS.</p>
https://doi.org/10.5281/zenodo.4147555
oai:zenodo.org:4147555
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147554
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
SModelS – a tool for interpreting simplified-model results from the LHC
info:eu-repo/semantics/lecture
oai:zenodo.org:4147361
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Adamec, Mat
2020-07-15
<p>Practical tutorial on how to perform an analysis with the Coffea set project packages.</p>
https://doi.org/10.5281/zenodo.4147361
oai:zenodo.org:4147361
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147360
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Columnar Analysis at Scale with Coffea
info:eu-repo/semantics/lecture
oai:zenodo.org:4147627
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Kropivnitskaya, Anna
2020-07-17
<p>The CMS inner tracking system is a fully silicon-based high precision detector. Accurate knowledge of the positions of active and inactive elements is important for simulating the detector, planning detector upgrades, and reconstructing charged particle tracks. Nuclear interactions of hadrons with the detector material create secondary vertices whose positions map the material with a sub-millimeter precision in situ, while the detector is collecting data from LHC collisions.</p>
<p>A neural network (NN) with two hidden layers was used to separate secondary vertices due to combinatorial background from those arising from nuclear interactions with material. The NN was trained and tested on data from proton-proton collisions at a center-of-mass energy of 13 TeV, recorded in 2018 at the LHC.</p>
<p>Supervised NN training is performed using Keras and Matplotlib in a Jupyter notebook. Secondary vertices in the training data are classified as signal or background, based on their geometrical position. Even though the variables used in training show only small differences between background and signal, the NN has impressive separation power. Tomographies of the CMS inner tracker detector before and after background cleaning are presented.</p>
<p> </p>
<p>Notebooks of the presentation.</p>
https://doi.org/10.5281/zenodo.4147627
oai:zenodo.org:4147627
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147626
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Machine learning technique for signal-background separation of nuclear interaction vertices in the CMS detector
info:eu-repo/semantics/lecture
oai:zenodo.org:4147611
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Schreiner, Henry Fredrick
Dembinski, Hans Peter
2020-07-17
<p>The boost-histogram library provides first-class histogram objects in Python. You can compose axes and a storage to fit almost any problem. You can fill, manipulate, slice, and project then, and pass them between other Scikit-HEP libraries like Uproot4, mplhep, and histoprint. Boost-histogram is meant to be the "NumPy" of histogram libraries that others can build on; the "pandas" of histograms is "Hist", a physicist friendly front-end that extends and expands boost-histogram to do plotting and more. An early version of Hist is shown for the first time here.</p>
https://doi.org/10.5281/zenodo.4147611
oai:zenodo.org:4147611
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147610
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
The boost-histogram package
info:eu-repo/semantics/lecture
oai:zenodo.org:4147609
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Novak, Andrzej
2020-07-17
<p>mplhep is small library on top of matplotlib, designed to simplify making typical HEP plots, which are not necessarily native to matplotlib, as well as, to distribute plotting styles and fonts to minimize the amount of needed cookie-cutter code and produce same results across platforms.</p>
https://doi.org/10.5281/zenodo.4147609
oai:zenodo.org:4147609
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147608
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
mplhep: bridging Matplotlib and HEP
info:eu-repo/semantics/lecture
oai:zenodo.org:3961646
2020-07-28T00:59:24Z
openaire
user-pyhep2020
Shyamsundar, Prasanth
2020-07-17
<p>Tutorial on the ThickBrick package for event selection and categorization in HEP.</p>
https://doi.org/10.5281/zenodo.3961646
oai:zenodo.org:3961646
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961645
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
ThickBrick: Optimal event selection and categorization in high energy physics
info:eu-repo/semantics/lecture
oai:zenodo.org:4147376
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Bonanomi, Matteo
2020-07-15
<p>In 2027 CERN is expected to start the High-Luminosity LHC (HL-LHC) phase. HL-LHC will integrate 10 times the current luminosity, leading to a high pile-up rate and unprecedented radiation levels. In order to cope with such a harsh environment and maintain the current physics performance, a major upgrade of the LHC detectors is required. As part of the HL-LHC detector upgrade programme, the CMS experiment is developing a High Granularity Calorimeter (HGCAL) to replace the existing endcap calorimeters.</p>
<p>Beam tests play a fundamental role in the validation of the detector design and in the study of its physics performance. In a typical beam test environment, it is important to have quick access to data, in order to perform explorative analysis, data visualization for the main physical distributions and to run data quality monitoring. During the offline analysis it is very often necessary to run comparisons between data and simulations, to identify problematic features in data, to develop preselection and cleaning cuts, and to reconstruct different observables distributions to produce the final results by means of statistical analysis.</p>
<p>Jupyter Notebooks are being used more and more by physicists to face this kind of task, as they provide an interactive interface where code can be executed, documented, and the outputs can be directly produced, analyzed and visualised. The data can be access remotely without the need of downloading to a local machine.</p>
<p>After a brief introduction about the HGCAL, the talk will focus on the major benefits that come from the use of interactive notebooks during a test beam campaign and in the subsequent phase of data analysis.</p>
<p>Notebook from the presentation.</p>
https://doi.org/10.5281/zenodo.4147376
oai:zenodo.org:4147376
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147375
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
High Granularity Calorimeter (HGCAL) test beam analysis using Jupyter notebooks
info:eu-repo/semantics/lecture
oai:zenodo.org:3961667
2020-07-28T00:59:24Z
openaire
user-pyhep2020
Kowalski, Jakub
Majewski, Maciej Witold
2020-07-13
<p>A detector monitoring framework for the vertex detector of the LHCb experiment.</p>
https://doi.org/10.5281/zenodo.3961667
oai:zenodo.org:3961667
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961666
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
TITANIA - how to structure detector monitoring
info:eu-repo/semantics/lecture
oai:zenodo.org:4147316
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Tejedor Saavedra, Enric
Wunsch, Stefan
2020-07-14
<p>Notebooks from the presentation on the latest and/or new PyROOT functionality.</p>
https://doi.org/10.5281/zenodo.4147316
oai:zenodo.org:4147316
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147315
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
A new PyROOT for ROOT 6.22
info:eu-repo/semantics/lecture
oai:zenodo.org:4147389
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Simpson, Nathan Daniel
2020-07-15
<p>The advent of deep learning has yielded powerful tools to automatically compute gradients of computations. This is because 'training a neural network' equates to iteratively updating its parameters using gradient descent to find the minimum of a loss function. Deep learning is then a subset of a broader paradigm; a workflow with free parameters that is end-to-end optimisable, provided one can keep track of the gradients all the way through. This paradigm is known as differentiable programming.</p>
<p>This work introduces neos, which is an example implementation of a fully differentiable HEP workflow, made possible by leveraging the Python modules <code>jax</code> and <code>pyhf</code>. In particular, through using a technique called fixed-point differentiation, neos makes the frequentist construction of the profile likelihood differentiable. This allows a neural network-based summary statistic to be trained with respect to the expected CLs value calculated downstream. Doing this results in an optimisation process that is aware of how every step in the workflow changes the CLs value, including the modelling and treatment of nuisance parameters.</p>
https://doi.org/10.5281/zenodo.4147389
oai:zenodo.org:4147389
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147388
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
neos: physics analysis as a differentiable program
info:eu-repo/semantics/lecture
oai:zenodo.org:3964086
2020-07-29T00:59:24Z
openaire
user-pyhep2020
Lust, Nate
2020-07-13
<p>Discussion of usage of Python at the Rubin Observatory.</p>
https://doi.org/10.5281/zenodo.3964086
oai:zenodo.org:3964086
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3964085
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Rubin Observatory: The software behind the science
info:eu-repo/semantics/lecture
oai:zenodo.org:3961663
2020-07-28T00:59:24Z
openaire
user-pyhep2020
Rodrigues, Eduardo
2020-07-13
<p>PyHEP 2020 welcome and workshop overview.</p>
https://doi.org/10.5281/zenodo.3961663
oai:zenodo.org:3961663
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961662
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Welcome and workshop overview
info:eu-repo/semantics/lecture
oai:zenodo.org:4147623
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Kropivnitskaya, Anna
2020-07-17
<p>The CMS inner tracking system is a fully silicon-based high precision detector. Accurate knowledge of the positions of active and inactive elements is important for simulating the detector, planning detector upgrades, and reconstructing charged particle tracks. Nuclear interactions of hadrons with the detector material create secondary vertices whose positions map the material with a sub-millimeter precision in situ, while the detector is collecting data from LHC collisions.</p>
<p>A neural network (NN) with two hidden layers was used to separate secondary vertices due to combinatorial background from those arising from nuclear interactions with material. The NN was trained and tested on data from proton-proton collisions at a center-of-mass energy of 13 TeV, recorded in 2018 at the LHC.</p>
<p>Supervised NN training is performed using Keras and Matplotlib in a Jupyter notebook. Secondary vertices in the training data are classified as signal or background, based on their geometrical position. Even though the variables used in training show only small differences between background and signal, the NN has impressive separation power. Tomographies of the CMS inner tracker detector before and after background cleaning are presented.</p>
<p>Slides of the presentation.</p>
https://doi.org/10.5281/zenodo.4147623
oai:zenodo.org:4147623
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147622
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Machine learning technique for signal-background separation of nuclear interaction vertices in the CMS detector
info:eu-repo/semantics/lecture
oai:zenodo.org:4152916
2020-10-31T12:27:04Z
openaire
user-pyhep2020
Matthew Feickert
2020-07-16
<p>The HistFactory p.d.f. template [<a href="https://cds.cern.ch/record/1456844">CERN-OPEN-2012-016</a>] is per-se independent of its implementation in ROOT and it is useful to be able to run statistical analysis outside of the ROOT, RooFit, RooStats framework. pyhf is a pure-Python implementation of that statistical model for multi-bin histogram-based analysis and its interval estimation is based on the asymptotic formulas of "Asymptotic formulae for likelihood-based tests of new physics" [<a href="https://arxiv.org/abs/1007.1727">arXiv:1007.1727</a>]. pyhf supports modern computational graph libraries such as TensorFlow and PyTorch in order to make use of features such as autodifferentiation and GPU acceleration, resulting in multiple times speedup of analysis fits. In addition, the JSON Schema developed to support the HistFactory specification has enabled for the open publication of four full likelihoods from ATLAS analyses to HEPData within a year of the first ever published.</p>
<p>In a tutorial the core features set of pyhf will be demonstrated in Jupyter notebooks in addition to how an analyst can use both the Python and command line APIs to quickly investigate a model and perform fits in an interactive session. The enabling features of the HistFactory JSON specification will also be explored.</p>
https://doi.org/10.5281/zenodo.4152916
oai:zenodo.org:4152916
eng
Zenodo
https://github.com/pyhf/tutorial-PyHEP-2020/tree/pyhep-2020
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4152915
info:eu-repo/semantics/openAccess
MIT License
https://opensource.org/licenses/MIT
PyHEP 2020 Workshop, 13-17 July 2020
pyhf
Scikit-HEP
physics
pyhf: Accelerating analyses and preserving likelihoods
info:eu-repo/semantics/lecture
oai:zenodo.org:3964090
2020-07-29T00:59:24Z
openaire
user-pyhep2020
Egede, Ulrik
2020-07-13
<p>Tutorial on how the job submission tool Ganga is exploiting virtualisation for user based large scale computations.</p>
https://doi.org/10.5281/zenodo.3964090
oai:zenodo.org:3964090
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3964089
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Ganga: flexible use of of virtualisation for user based large scale computations
info:eu-repo/semantics/lecture
oai:zenodo.org:4147433
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Dembinski, Hans Peter
2020-07-16
<p>Discussion of the iminuit fitting package - the package today and what's on the horizon.</p>
https://doi.org/10.5281/zenodo.4147433
oai:zenodo.org:4147433
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147432
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
iminuit: past and future
info:eu-repo/semantics/lecture
oai:zenodo.org:4147337
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Choi, Kyungeon
2020-07-14
<p>One of the biggest challenges in the HL-LHC era will be significantly more data to be recorded and analyzed from the collisions at the ATLAS and CMS detectors. ServiceX is a software R&D project in the area of Data Organization, Management and Access (DOMA) of the IRIS-HEP to investigate new analysis models for the HL-LHC era.</p>
<p>ServiceX is an experiment-agnostic service to enable on-demand data delivery specifically tailored for nearly-interactive vectorized analysis. It is capable of retrieving data from the grid, on-the-fly data transformation at the distributed system, and delivering user-selected data in a variety of different formats. The beauty of the service is coming from its modularity and diversity of transformation specified by a user in the query. ServiceX currently supports xAOD transformer for the ATLAS data and uproot transformer for the standard ROOT TTree. A transformer for the CMS data is under development.</p>
<p>The service is primarily written in Python language and utilizes several pythonic HEP tools such as uproot, Awkward Array, Functional ADL, and PyROOT in the backend of the service. The frontend library, fully written in Python, allows a user to make a query of data delivery even from a Jupyter notebook.</p>
<p>We will describe a harmonization of pythonic tools to achieve the functionalities delivered by ServiceX.</p>
https://doi.org/10.5281/zenodo.4147337
oai:zenodo.org:4147337
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147336
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
ServiceX: On-Demand Data Transformation and Delivery for the Present and HL-LHC Era
info:eu-repo/semantics/lecture
oai:zenodo.org:4136273
2020-10-27T00:26:55Z
openaire
user-pyhep2020
Shadura, Oksana
2020-07-15
<p>In the HL-LHC era, an order of magnitude increase of event rates will mean activities that can be done on a laptop today will require significantly more resources tomorrow. For example, increased dataset volumes means that users cannot necessarily keep all their data locally on a laptop - dedicated analysis facilities will be needed. Today, most facilities are batch-oriented while analysts often want to work interactive when exploring the data; facilities will likely need to provide a hybrid of both batch and interactive approaches going forward. U.S. CMS seeks to provide a prototype analysis facility that addresses these challenges during 2020. In this tutorial we describe and demonstrate elements of such a prototype at the University of Nebraska-Lincoln (UNL).</p>
<p>The prototype analysis facility provides services for “low latency columnar analysis”, enabling rapid processing of data in a column-wise fashion. These services, based on Dask and Jupyter notebooks, aim to dramatically lower time for analysis and provide an easily-scalable and user-friendly computational environment that will simplify, facilitate, and accelerate the delivery of HEP results. The facility is built on top of a local Kubernetes cluster and integrates dedicated resources with resources allocated via fairshare through the local HTCondor system. In addition to the user-facing interfaces such as Dask, the facility also manages access control through single-sign-on and authentication & authorization for data access. The showcase will include simple HEP analysis examples, managed interactively in a Jupyter notebook and scheduled on Dask workers and accessing both public and protected data</p>
https://doi.org/10.5281/zenodo.4136273
oai:zenodo.org:4136273
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4136272
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
A prototype U.S. CMS analysis facility
info:eu-repo/semantics/lecture
oai:zenodo.org:3961687
2020-07-28T00:59:24Z
openaire
user-pyhep2020
Pivarski, James
2020-07-13
<p>Tutorial on the uproot4 and awkward1 packages. for data manipulation in HEP.</p>
https://doi.org/10.5281/zenodo.3961687
oai:zenodo.org:3961687
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961686
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Uproot & Awkward Arrays
info:eu-repo/semantics/lecture
oai:zenodo.org:3961675
2020-07-28T10:09:53Z
openaire
user-pyhep2020
Pata, Joosep
2020-07-13
<p>Notebook tutorial showing how to connect <code>uproot</code>, <code>awkward-array</code>, <code>cupy</code> and <code>numba</code> to do numerically intensive physics data processing such as histogramming directly on a GPU.</p>
https://doi.org/10.5281/zenodo.3961675
oai:zenodo.org:3961675
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.3961674
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
Jagged physics data analysis with numba, awkward and uproot on a GPU
info:eu-repo/semantics/lecture
oai:zenodo.org:4147621
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Komm, Matthias
2020-07-17
<p>One of the most common data format in HEP are ROOT TTrees. However, modern machine learning packages require that the training data is supplied in different formats such as numpy arrays, HDF5, Pandas dataframes, or TFRecords instead. Hence, before training a machine learning algorithm, one usually has to first implement and perform a conversion and optional preprocessing of the training data set. Furthermore, if the data set is much larger than the available memory, dynamic loading and unloading of partial sets has to be performed efficiently during the training instead. To overcome both tasks, we developed a novel interface between ROOT and Tensorflow for training neural networks, where ROOT TTrees are streamed directly into the TensorFlow queue and threading system. This custom workflow allows a flexible selection of input features and the asynchronous preprocessing of data on the CPU in parallel to the neural network training on the GPU. The capabilities of the pipeline are demonstrated for training a classifier to identify displaced jets from the decay of new long-lived particles using a data sample of O(100 GB). Plans on future developments will be given as well.</p>
https://doi.org/10.5281/zenodo.4147621
oai:zenodo.org:4147621
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147620
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
ROOT preprocessing pipeline for machine learning with TensorFlow
info:eu-repo/semantics/lecture
oai:zenodo.org:4147312
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Tejedor Saavedra, Enric
Wunsch, Stefan
2020-07-14
<p>Slides from the presentation on the latest and/or new PyROOT functionality.</p>
https://doi.org/10.5281/zenodo.4147312
oai:zenodo.org:4147312
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147311
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
A new PyROOT for ROOT 6.22
info:eu-repo/semantics/lecture
oai:zenodo.org:4147328
2020-10-29T00:26:54Z
openaire
user-pyhep2020
Dembinski, Hans Peter
2020-07-14
<p>We present the resample package, which implements the resampling methods jackknife and Efron's bootstrap. Both are computationally intensive but highly general tools to compute biases and variances of estimators. The bootstrap allows one to compute the full covariance matrix for the output of an arbitrarily complex estimation problem, such as a multi-parameter fit, and confidence intervals for those parameters.</p>
https://doi.org/10.5281/zenodo.4147328
oai:zenodo.org:4147328
Zenodo
https://zenodo.org/communities/pyhep2020
https://doi.org/10.5281/zenodo.4147327
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
PyHEP 2020 Workshop, 13-17 July 2020
resample: use the bootstrap and jackknife from Python
info:eu-repo/semantics/lecture