Deep Learning techniques are being studied for different applications by the HEP community: in this talk, we discuss the case of detector simulation. The need for simulated events, expected in the ...future for LHC experiments and their High Luminosity upgrades, is increasing dramatically and requires new fast simulation solutions. Here we present updated results on the development of 3DGAN, one of the first examples using three-dimensional convolutional Generative Adversarial Networks to simulate high granularity electromagnetic calorimeters. In particular, we report on two main aspects: results on the simulation of a more general, realistic physics use case and on data parallel strategies to distribute the training process across multiple nodes on public cloud resources.
Particle physics is a branch of science aiming at discovering the fundamental laws of matter and forces. Graph neural networks are trainable functions which operate on graphs-sets of elements and ...their pairwise relations-and are a central method within the broader field of geometric deep learning. They are very expressive and have demonstrated superior performance to other classical deep learning approaches in a variety of domains. The data in particle physics are often represented by sets and graphs and as such, graph neural networks offer key advantages. Here we review various applications of graph neural networks in particle physics, including different graph constructions, model architectures and learning objectives, as well as key open problems in particle physics for which graph neural networks are promising.
The High Luminosity Large Hadron Collider (HL-LHC) at CERN will involve a significant increase in complexity and sheer size of data with respect to the current LHC experimental complex. Hence, the ...task of reconstructing the particle trajectories will become more involved due to the number of simultaneous collisions and the resulting increased detector occupancy. Aiming to identify the particle paths, machine learning techniques such as graph neural networks are being explored in the HEP.TrkX project and its successor, the Exa.TrkX project. Both show promising results and reduce the combinatorial nature of the problem. Previous results of our team have demonstrated the successful attempt of applying quantum graph neural networks to reconstruct the particle track based on the hits of the detector. A higher overall accuracy is gained by representing the training data in a meaningful way within an embedded space. That has been included in the Exa.TrkX project by applying a classical MLP. Consequently, pairs of hits belonging to different trajectories are pushed apart while those belonging to the same ones stay close together. We explore the applicability of variational quantum circuits that include a relatively low number of qubits applicable to NISQ devices within the task of embedding and show preliminary results.
Accurate determination of particle track reconstruction parameters will be a major challenge for the High Luminosity Large Hadron Collider (HL-LHC) experiments. The expected increase in the number of ...simultaneous collisions at the HL-LHC and the resulting high detector occupancy will make track reconstruction algorithms extremely demanding in terms of time and computing resources. The increase in number of hits will increase the complexity of track reconstruction algorithms. In addition, the ambiguity in assigning hits to particle tracks will be increased due to the finite resolution of the detector and the physical “closeness” of the hits. Thus, the reconstruction of charged particle tracks will be a major challenge to the correct interpretation of the HL-LHC data. Most methods currently in use are based on Kalman filters which are shown to be robust and to provide good physics performance. However, they are expected to scale worse than quadratically. Designing an algorithm capable of reducing the combinatorial background at the hit level, would provide a much “cleaner” initial seed to the Kalman filter, strongly reducing the total processing time. One of the salient features of Quantum Computers is the ability to evaluate a very large number of states simultaneously, making them an ideal instrument for searches in a large parameter space. In fact, different R&D initiatives are exploring how Quantum Tracking Algorithms could leverage such capabilities. In this paper, we present our work on the implementation of a quantum-based track finding algorithm aimed at reducing combinatorial background during the initial seeding stage. We use the publicly available dataset designed for the kaggle TrackML challenge.
The Computing Models of the LHC experiments continue to evolve from the simple hierarchical MONARC2 model towards more agile models where data is exchanged among many Tier2 and Tier3 sites, relying ...on both large scale file transfers with strategic data placement, and an increased use of remote access to object collections with caching through CMS's AAA, ATLAS' FAX and ALICE's AliEn projects, for example. The challenges presented by expanding needs for CPU, storage and network capacity as well as rapid handling of large datasets of file and object collections have pointed the way towards future more agile pervasive models that make best use of highly distributed heterogeneous resources. In this paper, we explore the use of Named Data Networking (NDN), a new Internet architecture focusing on content rather than the location of the data collections. As NDN has shown considerable promise in another data intensive field, Climate Science, we discuss the similarities and differences between the Climate and HEP use cases, along with specific issues HEP faces and will face during LHC Run2 and beyond, which NDN could address.
This paper reports on the second “Throughput” phase of the Tracking Machine Learning (TrackML) challenge on the Codalab platform. As in the first “Accuracy” phase, the participants had to solve a ...difficult experimental problem linked to tracking accurately the trajectory of particles as e.g. created at the Large Hadron Collider (LHC): given
O
(
10
5
)
points, the participants had to connect them into
O
(
10
4
)
individual groups that represent the particle trajectories which are approximated helical. While in the first phase only the accuracy mattered, the goal of this second phase was a compromise between the accuracy and the speed of inference. Both were measured on the Codalab platform where the participants had to upload their software. The best three participants had solutions with good accuracy and speed an order of magnitude faster than the state of the art when the challenge was designed. Although the core algorithms were less diverse than in the first phase, a diversity of techniques have been used and are described in this paper. The performance of the algorithms is analysed in depth and lessons derived.
The central Monte-Carlo production of the CMS experiment utilizes the WLCG infrastructure and manages daily thousands of tasks, each up to thousands of jobs. The distributed computing system is bound ...to sustain a certain rate of failures of various types, which are currently handled by computing operators a posteriori. Within the context of computing operations, and operation intelligence, we propose a Machine Learning technique to learn from the operators with a view to reduce the operational workload and delays. This work is in continuation of CMS work on operation intelligence to try and reach accurate predictions with Machine Learning. We present an approach to consider the log files of the workflows as regular text to leverage modern techniques from Natural Language Processing (NLP). In general, log files contain a substantial amount of text that is not human language. Therefore, different log parsing approaches are studied in order to map the log files’ words to high dimensional vectors. These vectors are then exploited as feature space to train a model that predicts the action that the operator has to take. This approach has the advantage that the information of the log files is extracted automatically and the format of the logs can be arbitrary. In this work the performance of the log file analysis with NLP is presented and compared to previous approaches.
In recent years, several studies have demonstrated the benefit of using deep learning to solve typical tasks related to high energy physics data taking and analysis. In particular, generative ...adversarial networks are a good candidate to supplement the simulation of the detector response in a collider environment. Training of neural network models has been made tractable with the improvement of optimization methods and the advent of GP-GPU well adapted to tackle the highly-parallelizable task of training neural nets. Despite these advancements, training of large models over large data sets can take days to weeks. Even more so, finding the best model architecture and settings can take many expensive trials. To get the best out of this new technology, it is important to scale up the available network-training resources and, consequently, to provide tools for optimal large-scale distributed training. In this context, our development of a new training workflow, which scales on multi-node/multi-GPU architectures with an eye to deployment on high performance computing machines is described. We describe the integration of hyper parameter optimization with a distributed training framework using Message Passing Interface, for models defined in keras 12 or pytorch 13. We present results on the speedup of training generative adversarial networks trained on a data set composed of the energy deposition from electron, photons, charged and neutral hadrons in a
fine grained digital calorimeter.
We discuss a model-independent strategy for boosting new physics searches with the help of an unsupervised anomaly detection algorithm. Prior to a search, each input event is preprocessed by the ...algorithm - a variational autoencoder (VAE). Based on the loss assigned to each event, input data can be split into a background control sample and a signal enriched sample. Following this strategy, one can enhance the sensitivity to new physics with no assumption on the underlying new physics signature. Our results show that a typical BSM search on the signal enriched group is more sensitive than an equivalent search on the original dataset.
The certification of the CMS experiment data as usable for physics analysis is a crucial task to ensure the quality of all physics results published by the collaboration. Currently, the certification ...conducted by human experts is labor intensive and based on the scrutiny of distributions integrated on several hours of data taking. This contribution focuses on the design and prototype of an automated certification system assessing data quality on a per-luminosity section (i.e. 23 seconds of data taking) basis. Anomalies caused by detector malfunctioning or sub-optimal reconstruction are difficult to enumerate a priori and occur rarely, making it difficult to use classical supervised classification methods such as feedforward neural networks. We base our prototype on a semi-supervised approach which employs deep autoencoders. This approach has been qualified successfully on CMS data collected during the 2016 LHC run: we demonstrate its ability to detect anomalies with high accuracy and low false positive rate, when compared against the outcome of the manual certification by experts. A key advantage of this approach over other machine learning technologies is the great interpretability of the results, which can be further used to ascribe the origin of the problems in the data to a specific sub-detector or physics objects.