A5055650771- Particle physics theoretical and experimental studies
- High-Energy Particle Collisions Research
- Particle Detector Development and Performance
- Quantum Chromodynamics and Particle Interactions
- Dark Matter and Cosmic Phenomena
- Computational Physics and Python Applications
- Neutrino Physics Research
- Cosmology and Gravitation Theories
- Radiomics and Machine Learning in Medical Imaging
- Radiation Detection and Scintillator Technologies
- Distributed and Parallel Computing Systems
- Medical Imaging Techniques and Applications
- Pulsars and Gravitational Waves Research
- Gaussian Processes and Bayesian Inference
- Advanced Data Storage Technologies
- Astrophysics and Cosmic Phenomena
- Particle Accelerators and Free-Electron Lasers
- Atomic and Subatomic Physics Research
- Superconducting Materials and Applications
- Generative Adversarial Networks and Image Synthesis
- Big Data Technologies and Applications
- Black Holes and Theoretical Physics
- Statistical Methods and Inference
- Gamma-ray bursts and supernovae
- Scientific Computing and Data Management
University of Arizona
2022-2025
University of California, Irvine
2003-2025
Lawrence Berkeley National Laboratory
2021-2025
University of California System
2025
Université Paris-Saclay
2019-2024
Laboratoire de Physique des 2 Infinis Irène Joliot-Curie
2019-2024
Institut National de Physique Nucléaire et de Physique des Particules
2020-2024
Centre National de la Recherche Scientifique
2019-2024
University of Iowa
2019-2024
Laboratoire Leprince-Ringuet
2024
Abstract Neutron stars provide a unique laboratory for studying matter at extreme pressures and densities. While there is no direct way to explore their interior structure, X-rays emitted from these can indirectly clues the equation of state (EOS) superdense nuclear through inference star's mass radius. However, EOS directly X-ray spectra extremely challenging complicated by systematic uncertainties. The current art use simulation-based likelihoods in piece-wise method which relies on...
Machine learning techniques are becoming an integral component of data analysis in high energy physics. These tools provide a significant improvement sensitivity over traditional analyses by exploiting subtle patterns high-dimensional feature spaces. may not be well modeled the simulations used for training machine methods, resulting enhanced to systematic uncertainties. Contrary wisdom constructing strategy that is invariant uncertainties, we study use classifier fully aware uncertainties...
Abstract A variety of techniques have been proposed to train machine learning classifiers that are independent a given feature. While this can be an essential technique for enabling background estimation, it may also useful reducing uncertainties. We carefully examine theory uncertainties, which typically do not statistical origin. will provide explicit examples two-point (fragmentation modeling) and continuous (higher-order corrections) uncertainties where decorrelating significantly...
Abstract Neutron stars provide a unique opportunity to study strongly interacting matter under extreme density conditions. The intricacies of inside neutron and their equation state are not directly visible, but determine bulk properties, such as mass radius, which affect the star's thermal X-ray emissions. However, telescope spectra these emissions also affected by stellar distance, hydrogen column, effective surface temperature, always well-constrained. Uncertainties on nuisance parameters...
Detectors of High Energy Physics experiments, such as the ATLAS dectector [1] at Large Hadron Collider [2], serve cameras that take pictures particles produced in collision events. One key detector technologies used for measuring energy are calorimeters. Particles will lose their a cascade (called shower) electromagnetic and hadronic interactions with dense absorbing material. The number this showering process is subsequently measured across sampling layers calorimeter. deposition...
High-energy collisions at the Large Hadron Collider (LHC) provide valuable insights into open questions in particle physics. However, detector effects must be corrected before measurements can compared to certain theoretical predictions or from other detectors. Methods solve this \textit{inverse problem} of mapping observations quantities underlying collision are essential parts many physics analyses LHC. We investigate and compare various generative deep learning methods approximate inverse...
The measurements performed by particle physics experiments must account for the imperfect response of detectors used to observe interactions. One approach, unfolding, statistically adjusts experimental data detector effects. Recently, generative machine learning models have shown promise performing unbinned unfolding in a high number dimensions. However, all current approaches are limited fixed set observables, making them unable perform full-event variable dimensional environment collider...
Recognizing symmetries in data allows for significant boosts neural network training, which is especially important where training are limited. In many cases, however, the exact underlying symmetry present only an idealized dataset, and broken actual data, due to asymmetries detector, or varying response resolution as a function of particle momentum. Standard approaches, such augmentation equivariant networks fail represent nature full, symmetry, effectively overconstraining network. We...
Hadronization is a complex quantum process whereby quarks and gluons become hadrons. The widely-used models of hadronization in event generators are based on physically-inspired phenomenological with many free parameters. We propose an alternative approach neural networks used instead. Deep generative highly flexible, differentiable, compatible Graphical Processing Unit (GPUs). make the first step towards data-driven machine learning-based model by replacing compont within Herwig generator...
Abstract Generation of simulated detector response to collision products is crucial data analysis in particle physics, but computationally very expensive. One subdetector, the calorimeter, dominates computational time due high granularity its cells and complexity interactions. Generative models can provide more rapid sample production, currently require significant effort optimize performance for specific geometries, often requiring many describe varying cell sizes arrangements, without...
Abstract The need for large scale and high fidelity simulated samples the ATLAS experiment motivates development of new simulation techniques. Building on recent success deep learning algorithms at interpolation as well image generation, Variational Auto-Encoders Generative Adversarial Networks are investigated modeling response electromagnetic calorimeter photons in a central region over range energies. synthesized showers compared to from full detector using Geant4. This study demonstrates...
The increase of the particle flux at HL-LHC with instantaneous luminosities up to L=7.5 × 10 34 cm −2 s −1 will have a severe impact on ATLAS detector reconstruction and trigger performance. end-cap forward region where liquid Argon calorimeter has coarser granularity inner tracker poorer momentum resolution be particularly affected. A High Granularity Timing Detector installed in front calorimeters help charged-particle luminosity measurement. This low angle is introduced augment new...
A comprehensive uncertainty estimation is vital for the precision program of LHC. While experimental uncertainties are often described by stochastic processes and well-defined nuisance parameters, theoretical lack such a description. We study estimates cross-section predictions based on scale variations across large set processes. find patterns similar to origin, with accurate mediated strong force, but systematic underestimate electroweak propose an improved scheme, variation reference...
Abstract The interiors of neutron stars reach densities and temperatures beyond the limits terrestrial experiments, providing vital laboratories for probing nuclear physics. While star's interior is not directly observable, its pressure density determine macroscopic structure which affects spectra observed in telescopes. relationship between observations internal state complex partially intractable, presenting difficulties inference. Previous work has focused on regression from stellar...
The measurements performed by particle physics experiments must account for the imperfect response of detectors used to observe interactions. One approach, unfolding, statistically adjusts experimental data detector effects. Recently, generative machine learning models have shown promise performing unbinned unfolding in a high number dimensions. However, all current approaches are limited fixed set observables, making them unable perform full-event variable dimensional environment collider...
We report results of a search for nuclear recoils induced by weakly interacting massive particle (WIMP) dark matter using the LUX-ZEPLIN (LZ) two-phase xenon time projection chamber. This analysis uses total exposure $4.2\pm0.1$ tonne-years from 280 live days LZ operation, which $3.3\pm0.1$ and 220 are new. A technique to actively tag background electronic $^{214}$Pb $\beta$ decays is featured first time. Enhanced electron-ion recombination observed in two-neutrino double electron capture...
Embedded microprocessor cores are increasingly being used in embedded and mobile devices. The software running on these is often a priori known, thus, there an opportunity for customizing the cache subsystem improved performance. In this work, we propose efficient algorithm to directly compute parameters satisfying desired performance criteria. Our approach avoids simulation exhaustive exploration, and, instead, relies exact algorithmic approach. We demonstrate feasibility of our by applying...
Generation of simulated detector response to collision products is crucial data analysis in particle physics, but computationally very expensive. One subdetector, the calorimeter, dominates computational time due high granularity its cells and complexity interactions. Generative models can provide more rapid sample production, currently require significant effort optimize performance for specific geometries, often requiring many describe varying cell sizes arrangements, without ability...
Estimating uncertainty is at the core of performing scientific measurements in HEP: a measurement not useful without an estimate its uncertainty. The goal quantification (UQ) inextricably linked to question, "how do we physically and statistically interpret these uncertainties?" answer this question depends only on computational task aim undertake, but also methods use for that task. For artificial intelligence (AI) applications HEP, there are several areas where interpretable UQ essential,...
Calorimeter shower simulations are often the bottleneck in simulation time for particle physics detectors. A lot of effort is currently spent on optimizing generative architectures specific detector geometries, which generalize poorly. We develop a geometry-aware autoregressive model range calorimeter geometries such that learns to adapt its energy deposition depending size and position cells. This key proof-of-concept step towards building can new unseen with little no additional training....