A5005322378- Particle physics theoretical and experimental studies
- Particle Detector Development and Performance
- Computational Physics and Python Applications
- Quantum Chromodynamics and Particle Interactions
- Black Holes and Theoretical Physics
- Dark Matter and Cosmic Phenomena
- High-Energy Particle Collisions Research
- Cosmology and Gravitation Theories
- Radiation Detection and Scintillator Technologies
- Distributed and Parallel Computing Systems
- Gaussian Processes and Bayesian Inference
- Neutrino Physics Research
- Astrophysics and Cosmic Phenomena
- Simulation Techniques and Applications
- Generative Adversarial Networks and Image Synthesis
- Soil Moisture and Remote Sensing
- Advanced Data Storage Technologies
- Speech Recognition and Synthesis
- Nuclear reactor physics and engineering
- Radio Astronomy Observations and Technology
- Electron and X-Ray Spectroscopy Techniques
- Medical Imaging Techniques and Applications
- Radiation Therapy and Dosimetry
- Scientific Research and Discoveries
- Noncommutative and Quantum Gravity Theories
Austrian Academy of Sciences
2023-2025
Heidelberg University
2022-2025
Rutgers, The State University of New Jersey
2021-2024
Rutgers Sexual and Reproductive Health and Rights
2021-2024
Institute of High Energy Physics
2023-2024
HES-SO Genève
2023
Deutsches Elektronen-Synchrotron DESY
2023
Universität Hamburg
2023
Heidelberg University
2023
University of Geneva
2023
We present a novel integrator based on normalizing flows which can be used to improve the unweighting efficiency of Monte-Carlo event generators for collider physics simulations. In contrast machine learning approaches surrogate models, our method generates correct result even if underlying neural networks are not optimally trained. exemplify new strategy using example Drell-Yan type processes at LHC, both leading and partially next-to-leading order QCD.
We propose a new model-agnostic search strategy for physics beyond the standard model (BSM) at LHC, based on novel application of neural density estimation to anomaly detection. Our approach, which we call Classifying Anomalies THrough Outer Density Estimation (CATHODE), assumes BSM signal is localized in region (defined e.g. using invariant mass). By training conditional estimator collection additional features outside region, interpolating it into and sampling from it, produce events that...
First-principle simulations are at the heart of high-energy physics research program. They link vast data output multi-purpose detectors with fundamental theory predictions and interpretation. This review illustrates a wide range applications modern machine learning to event generation simulation-based inference, including conceptional developments driven by specific requirements particle physics. New ideas tools developed interface will improve speed precision forward simulations, handle...
We introduce CaloFlow, a fast detector simulation framework based on normalizing flows. For the first time, we demonstrate that flows can reproduce many-channel calorimeter showers with extremely high fidelity, providing fresh alternative to computationally expensive GEANT4 simulations, as well other state-of-the-art frameworks GANs and VAEs. Besides usual histograms of physical features images showers, new metric for judging quality generative modeling: performance classifier trained...
We provide an overview of the status Monte-Carlo event generators for high-energy particle physics. Guided by experimental needs and requirements, we highlight areas active development, opportunities future improvements. Particular emphasis is given to physics models algorithms that are employed across a variety experiments. These common themes in generator development lead more comprehensive understanding at highest energies intensities, allow be tested against wealth data have been...
Well-trained classifiers and their complete weight distributions provide us with a well-motivated practicable method to test generative networks in particle physics. We illustrate benefits for distribution-shifted jets, calorimeter showers, reconstruction-level events. In all cases, the classifier weights make powerful of network, identify potential problems density estimation, relate them underlying physics, tie comprehensive precision uncertainty treatment networks.
Simulating particle detector response is the single most expensive step in Large Hadron Collider computational pipeline. Recently it was shown that normalizing flows can accelerate this process while achieving unprecedented levels of accuracy, but scaling approach up to higher resolutions relevant for future upgrades leads prohibitive memory constraints. To overcome problem, we introduce Inductive CaloFlow (icaloflow), a framework fast simulation based on an inductive series trained pattern...
Abstract Machine learning-based anomaly detection (AD) methods are promising tools for extending the coverage of searches physics beyond Standard Model (BSM). One class AD that has received significant attention is resonant detection, where BSM assumed to be localized in at least one known variable. While there have been many proposed identify such a signal make use simulated or detected data different ways, not yet study methods’ complementarity. To this end, we address two questions....
CALOFLOW is a new and promising approach to fast calorimeter simulation based on normalizing flows. Applying the photon charged pion ≥ant showers of Dataset 1 Fast Calorimeter Simulation Challenge 2022, we show how it can produce high-fidelity samples with sampling time that several orders magnitude faster than ≥ant. We demonstrate fidelity using shower images, histograms high level features, aggregate metrics such as classifier trained distinguish from samples.
We discuss the systematics of power counting in general effective field theories, focusing on those that are nonrenormalizable at leading order. As an illuminating example we consider chiral perturbation theory gauged under electromagnetic U(1) symmetry. This describes low-energy interactions octet pseudo-Goldstone bosons QCD with photons and has been discussed extensively literature. Peculiarities standard approach pointed out it is shown how these resolved within our scheme. The...
We provide an overview of the status Monte-Carlo event generators for high-energy particle physics. Guided by experimental needs and requirements, we highlight areas active development, opportunities future improvements. Particular emphasis is given to physics models algorithms that are employed across a variety experiments. These common themes in generator development lead more comprehensive understanding at highest energies intensities, allow be tested against wealth data have been...
Recently, we introduced caloflow, a high-fidelity generative model for geant4 calorimeter shower emulation based on normalizing flows. Here, present caloflow v2, an improvement our original framework that speeds up generation by further factor of 500 relative to the original. The is technique called probability density distillation, originally developed speech synthesis in machine learning literature, and which develop introducing set powerful new loss terms. We demonstrate v2 preserves same...
Abstract We explore the use of normalizing flows to emulate Monte Carlo detector simulations photon showers in a high-granularity electromagnetic calorimeter prototype for International Large Detector (ILD). Our proposed method — which we refer as “Layer-to-Layer Flows” ( L2LFlows ) is an evolution CaloFlow architecture adapted higher-dimensional setting (30 layers 10× 10 voxels each). The main innovation consists introducing 30 separate flows, one each layer calorimeter, where flow...
The full optimization of the design and operation instruments whose functioning relies on interaction radiation with matter is a super-human task, due to large dimensionality space possible choices for geometry, detection technology, materials, data-acquisition, information-extraction techniques, interdependence related parameters. On other hand, massive potential gains in performance over standard, "experience-driven" layouts are principle within our reach if an objective function fully...
In many well-motivated models of the electroweak scale, cascade decays new particles can result in highly boosted hadronic resonances (e.g., <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" display="inline"><a:mi>Z</a:mi><a:mo>/</a:mo><a:mi>W</a:mi><a:mo>/</a:mo><a:mi>h</a:mi></a:math>). This make these rich and promising targets for recently developed resonant anomaly detection methods powered by modern machine learning. We demonstrate this using state-of-the-art classifying anomalies...
Whenever invertible generative networks are needed for LHC physics, normalizing flows show excellent performance. In this work, we investigate their performance fast calorimeter shower simulations with increasing phase space dimension. We use and expressive coupling spline transformations applied to the CaloChallenge datasets. addition base flow architecture also employ a VAE compress dimensionality train network in latent space. evaluate our on several metrics, including high-level...
We consider the electroweak chiral Lagrangian, including a light scalar boson, in limit of small ξ=v2/f2. Here v is scale and f corresponding new strong dynamics. show how conventional SILH defined as effective theory strongly-interacting Higgs (SILH) to first order ξ, can be obtained limiting case complete Lagrangian. The approach presented here ensures completeness operator basis at considered order, it clarifies systematics guarantees consistent unambiguous power counting, shows...
We perform a Bayesian statistical analysis of the constraints on nonlinear Effective Theory given by Higgs electroweak chiral Lagrangian. obtain bounds effective coefficients entering in observables at leading order, using all available Higgs-boson signal strengths from LHC runs 1 and 2. Using prior dependence study solutions, we discuss results within context natural-sized Wilson coefficients. further expected sensitivities to different various possible future colliders. Finally, interpret...
Theory predictions for the LHC require precise numerical phase-space integration and generation of unweighted events. We combine machine-learned multi-channel weights with a normalizing flow importance sampling, to improve classical methods integration. develop an efficient bi-directional setup based on invertible network, combining online buffered training potentially expensive integrands. illustrate our method Drell-Yan process additional narrow resonance.
In a recent paper we showed that the electroweak chiral Lagrangian at leading order is equivalent to conventional $$\kappa $$ formalism used by ATLAS and CMS test Higgs anomalous couplings. Here apply this fact fit latest data. The new aspect of our analysis systematic interpretation parameters within an EFT. Concentrating on processes production decay have been measured so far, six turn out be relevant: $$c_V$$ , $$c_t$$ $$c_b$$ $$c_\tau $$c_{\gamma \gamma }$$ $$c_{gg}$$ . A global Bayesian...
If the electroweak phase transition (EWPT) is of strongly first order due to higher dimensional operators, scale new physics generating them at TeV or below. In this case effective-field theory (EFT) neglecting operators dimension than six may overlook terms that are relevant for EWPT analysis. article we study in EFT eight. We estimate reach future gravitational wave observatory LISA probing region which and compare it with capabilities Higgs measurements via double-Higgs production current...
We propose a parametrization of anomalous Higgs-boson couplings that is both systematic and practical. It based on the electroweak chiral Lagrangian, including light Higgs boson, as effective field theory (EFT) at scale v. This appropriate framework for case sizeable deviations in order 10% from Standard Model, considered to be parametrically larger than new-physics effects sector gauge interactions. The role power counting identifying relevant parameters emphasized. three scales, v, new...