A5046520638- Scientific Computing and Data Management
- Advanced Data Storage Technologies
- Distributed and Parallel Computing Systems
- Galaxies: Formation, Evolution, Phenomena
- Astronomy and Astrophysical Research
- Adaptive optics and wavefront sensing
- Gamma-ray bursts and supernovae
- Computational Physics and Python Applications
- Parallel Computing and Optimization Techniques
- Cosmology and Gravitation Theories
- Cloud Computing and Resource Management
- Research Data Management Practices
- Particle physics theoretical and experimental studies
- Advanced X-ray Imaging Techniques
- Advanced Image Processing Techniques
- Generative Adversarial Networks and Image Synthesis
- Enzyme Structure and Function
- Astrophysics and Cosmic Phenomena
- Astronomical Observations and Instrumentation
- Statistical and numerical algorithms
- Gaussian Processes and Bayesian Inference
- History and Developments in Astronomy
- Information and Cyber Security
- Magnetic confinement fusion research
- Knowledge Management and Technology
Lawrence Berkeley National Laboratory
2016-2024
National Energy Research Scientific Computing Center
2016-2024
SLAC National Accelerator Laboratory
2012-2019
Argonne National Laboratory
2017
Kavli Institute for Particle Astrophysics and Cosmology
2012-2016
Menlo School
2016
Stanford University
2012-2015
(Abridged) We describe here the most ambitious survey currently planned in optical, Large Synoptic Survey Telescope (LSST). A vast array of science will be enabled by a single wide-deep-fast sky survey, and LSST have unique capability faint time domain. The design is driven four main themes: probing dark energy matter, taking an inventory Solar System, exploring transient optical sky, mapping Milky Way. wide-field ground-based system sited at Cerro Pach\'{o}n northern Chile. telescope 8.4 m...
We present first results from the third GRavitational lEnsing Accuracy Testing (GREAT3) challenge, in a sequence of challenges for testing methods inferring weak gravitational lensing shear distortions simulated galaxy images. GREAT3 was divided into experiments to test three specific questions, and included space- ground-based data with constant or cosmologically varying fields. The simplest (control) experiment parametric galaxies realistic distribution signal-to-noise, size, ellipticity,...
Deep learning is a promising tool to determine the physical model that describes our universe. To handle considerable computational cost of this problem, we present CosmoFlow: highly scalable deep application built on top TensorFlow framework. CosmoFlow uses efficient implementations 3D convolution and pooling primitives, together with improvements in threading for many element-wise operations, improve training performance Intel <sup xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Abstract Inferring model parameters from experimental data is a grand challenge in many sciences, including cosmology. This often relies critically on high fidelity numerical simulations, which are prohibitively computationally expensive. The application of deep learning techniques to generative modeling renewing interest using dimensional density estimators as inexpensive emulators fully-fledged simulations. These models have the potential make dramatic shift field scientific but for that...
Summary X‐ray scattering experiments using free electron lasers (XFELs) are a powerful tool to determine the molecular structure and function of unknown samples (such as COVID‐19 viral proteins). XFEL challenge computing in two ways: (i) due high cost running XFELs, fast turnaround time from data acquisition analysis is essential make informed decisions on experimental protocols; (ii) data‐collection rates growing exponentially, requiring new scalable algorithms. Here we report our...
We address the increasing complexity of scientific workflows in context high-performance computing (HPC) and their associated need for robust, adaptable, flexible computational support systems. explore five key trends as well future challenges opportunities HPC technologies.
Point estimators for the shearing of galaxy images induced by gravitational lensing involve a complex inverse problem in presence noise, pixelization, and model uncertainties. We present probabilistic forward modeling approach to inference that has potential mitigate biased inferences most common point is practical upcoming surveys. The first part our statistical framework requires specification likelihood function pixel data an imaging survey given parameterized models galaxies images....
The statistics of peak counts in reconstructed shear maps contain information beyond the power spectrum, and can improve cosmological constraints from measurements spectrum alone if systematic errors be controlled. We study effect galaxy shape measurement on predicted with Large Synoptic Survey Telescope (LSST). use LSST image simulator combination N-body simulations to model realistic for different models. include both noise and, first time, shapes. find that considered have relatively...
The complete 10-year survey from the Large Synoptic Survey Telescope (LSST) will image $\sim$ 20,000 square degrees of sky in six filter bands every few nights, bringing final depth to $r\sim27.5$, with over 4 billion well measured galaxies. To take full advantage this unprecedented statistical power, systematic errors associated weak lensing measurements need be controlled a level similar errors. This work is first attempt quantitatively estimate absolute and properties on shear due most...
A main science goal for the Large Synoptic Survey Telescope (LSST) is to measure cosmic shear signal from weak lensing extreme accuracy. One difficulty, however, that with short exposure time (≃15 s) proposed, spatial variation of point spread function (PSF) shapes may be dominated by atmosphere, in addition optics errors. While errors mainly cause PSF vary on angular scales similar or larger than a single CCD sensor, atmosphere generates stochastic structures wide range scales. It thus...
Initial studies have suggested generative adversarial networks (GANs) promise as fast simulations within HEP. These studies, while promising, been insufficiently precise and also, like GANs in general, suffer from stability issues.We apply to generate full particle physics events (not individual objects), explore conditioning of generated based on theory parameters evaluate the precision generalization produced datasets. We this SUSY mass parameter interpolation pileup generation. also...
Experimental and observational instruments for scientific research (such as light sources, genome sequencers, accelerators, telescopes electron microscopes) increasingly require High Performance Computing (HPC) scale capabilities data analysis workflow processing. Next-generation are being deployed with higher resolutions faster capture rates, creating a big crunch that cannot be handled by modest institutional computing resources. Often these pipelines also near real-time have resilience...
In January 2019, the US Department of Energy, Office Science program in Advanced Scientific Computing Research, convened a workshop to identify priority research directions (PRDs) for situ data management (ISDM). A fundamental finding this is that methodologies used manage among variety tasks can be facilitate scientific discovery from many different sources—simulation, experiment, and sensors, example—and being able do so at numerous computing scales will benefit real-time decision-making,...
The nature of dark energy and the complete theory gravity are two central questions currently facing cosmology. A vital tool for addressing them is 3-point correlation function (3PCF), which probes deviations from a spatially random distribution galaxies. However, 3PCF's formidable computational expense has prevented its application to astronomical surveys comprising millions billions We present Galactos, high-performance implementation novel, O(N2) algorithm that uses load-balanced k-d tree...