检索结果-内蒙古大学图书馆

Point cloud-based diffusion models for the Electron-Ion Collider

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Araz, Jack Y. Mikuni, Vinicius Ringer, Felix Sato, Nobuo Acosta, Fernando Torales Whitehill, Richard Center for Nuclear Theory Department of Physics and Astronomy Stony Brook University New York11794 United States Thomas Jefferson National Accelerator Facility Newport NewsVA23606 United States Department of Physics Old Dominion University NorfolkVA23529 United States National Energy Research Scientific Computing Center Berkeley Lab BerkeleyCA94720 United States Physics Division Lawrence Berkeley National Laboratory BerkeleyCA94720 United States

At high-energy collider experiments, generative models can be used for a wide range of tasks, including fast detector simulations, unfolding, searches of physics beyond the Standard Model, and inference tasks. In particular, it has been demonstrated that score-based diffusion models can generate high-fidelity and accurate samples of jets or collider events. This work expands on previous generative models in three distinct ways. First, our model is trained to generate entire collider events, including all particle species with complete kinematic information. We quantify how well the model learns event-wide constraints such as the conservation of momentum and discrete quantum numbers. We focus on the events at the future Electron-Ion Collider, but we expect that our results can be extended to proton-proton and heavy-ion collisions. Second, previous generative models often relied on image-based techniques. The sparsity of the data can negatively affect the fidelity and sampling time of the model. We address these issues using point clouds and a novel architecture combining edge creation with transformer modules called Point Edge Transformers. Third, we adapt the foundation model OmniLearn, to generate full collider events. This approach may indicate a transition toward adapting and fine-tuning foundation models for downstream tasks instead of training new models from scratch. © 2024, CC BY.

关键词： Heavy ions

Gravitational-wave searches for cosmic string cusps in Einstein Telescope data using deep learning

学校读者我要写书评

暂无评论

Physical Review D 2024年第2期109卷 022006-022006页

作者： Quirijn Meijer Melissa Lopez Daichi Tsuna Sarah Caudill Institute for Gravitational and Subatomic Physics (GRASP) Department of Physics Utrecht University Princetonplein 1 3584CC Utrecht The Netherlands Nikhef Science Park 105 1098XG Amsterdam The Netherlands TAPIR Mailcode 350-17 California Institute of Technology Pasadena California 91125 USA Research Center for the Early Universe (RESCEU) School of Science The University of Tokyo 7-3-1 Hongo Bunkyo-ku Tokyo 113-0033 Japan Department of Physics University of Massachusetts Dartmouth Massachusetts 02747 USA Center for Scientific Computing and Data Science Research University of Massachusetts Dartmouth Massachusetts 02747 USA

Gravitational-wave searches for cosmic strings are currently hindered by the presence of detector glitches, some classes of which strongly resemble cosmic string signals. This confusion greatly reduces the efficiency of searches. A deep-learning model is proposed for the task of distinguishing between gravitational-wave signals from cosmic string cusps and simulated blip glitches in design sensitivity data from the future Einstein Telescope. The model is an ensemble consisting of three convolutional neural networks, achieving an accuracy of 79%, a true positive rate of 76%, and a false positive rate of 18%. This marks the first time convolutional neural networks have been trained on a realistic population of Einstein Telescope glitches. On a dataset consisting of signals and glitches, the model is shown to outperform matched filtering, specifically being better at rejecting glitches. The behaviour of the model is interpreted through the application of several methods, including a novel technique called waveform surgery, used to quantify the importance of waveform sections to a classification model. In addition, a method to visualize convolutional neural network activations for one-dimensional time series is proposed and used. These analyses help further the understanding of the morphological differences between cosmic string cusp signals and blip glitches. Because of its classification speed in the order of magnitude of milliseconds, the deep-learning model is suitable for future use as part of a real-time detection pipeline. The deep-learning model is transverse and can therefore potentially be applied to other transient searches.

关键词： Cosmic strings & domain walls General relativity Gravitational waves

Fully precessing higher-mode surrogate model of effective-one-body waveforms

学校读者我要写书评

暂无评论

Physical Review D 2024年第12期110卷 124038-124038页

作者： Bhooshan Gadre Michael Pürrer Scott E. Field Serguei Ossokine Vijay Varma Max Planck Institute for Gravitational Physics (Albert Einstein Institute) Am Mühlenberg 1 Potsdam 14476 Germany Institute for Gravitational and Subatomic Physics (GRASP) Utrecht University Princetonplein 1 3584 CC Utrecht The Netherlands Department of Physics East Hall University of Rhode Island Kingston Rhode Island 02881 USA URI Center for Computational Research Tyler Hall University of Rhode Island Kingston Rhode Island 02881 USA Department of Mathematics University of Massachusetts Dartmouth Massachusetts 02747 USA Center for Scientific Computing and Data Science Research University of Massachusetts Dartmouth Massachusetts 02747 USA

We present a surrogate model of SEOBNRv4PHM, a fully precessing time-domain effective-one-body waveform model including subdominant modes. We follow an approach similar to that used to build recent numerical relativity surrogate models. Our surrogate is 5000 M in duration, covers mass ratios up to 1∶20 and dimensionless spin magnitudes up to 0.8. Validating the surrogate against an independent test set, we find that the median mismatch error is less than 10−3, which is typically smaller than the modeling error of SEOBNRv4PHM itself. At high total mass, a few percent of configurations can exceed mismatches of 10−2 if they are highly precessing and exceed a mass ratio of 1∶4. This surrogate is nearly 2 orders of magnitude faster than the underlying time-domain SEOBNRv4PHM model and can be evaluated in ∼50 ms. Bayesian inference analyses with SEOBNRv4PHM are typically very computationally demanding and can take from weeks to months to complete. The 2 order of magnitude speedup attained by our surrogate model enables practical parameter estimation analyses with this waveform family. This is crucial because Bayesian inference allows us to recover the masses and spins of binary black hole mergers given a model of the emitted gravitational waveform along with a description of the noise.

关键词：

The Case for Non-Volatile RAM in Cloud HPCaaS

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Fridman, Yehonatan Harel, Re'em Oren, Gal Department of Computer Science Ben-Gurion University of the Negev Israel Israel Atomic Energy Commission Scientific Computing Center Nuclear Research Center Negev Israel Department of Physics Nuclear Research Center Negev Israel Department of Computer Science Technion Israel Institute of Technology Israel

HPC as a service (HPCaaS) is a new way to expose HPC resources via cloud services. However, continued effort to port large-scale tightly coupled applications with high interprocessor communication to multiple (and many) nodes synchronously, as in on-premise supercomputers, is still far from satisfactory due to network latencies. As a consequence, in said cases, HPCaaS is recommended to be used with one or few instances. In this paper we take the claim that new piece of memory hardware, namely Non-Volatile RAM (NVRAM), can allow such computations to scale up to an order of magnitude with marginalized penalty in comparison to RAM. Moreover, we suggest that the introduction of NVRAM to HPCaaS can be cost-effective to the users and the suppliers in numerous forms. © 2022, CC BY.

关键词： Random access storage

Improved analysis of GW190412 with a precessing numerical relativity surrogate waveform model

学校读者我要写书评

暂无评论

Physical Review D 2021年第10期103卷 104027-104027页

作者： Tousif Islam Scott E. Field Carl-Johan Haster Rory Smith Department of Physics and Center for Scientific Computing and Visualization Research University of Massachusetts Dartmouth Massachusetts 02747 USA Department of Mathematics and Center for Scientific Computing and Visualization Research University of Massachusetts Dartmouth Massachusetts 02747 USA LIGO Laboratory Massachusetts Institute of Technology 185 Albany St Cambridge Massachusetts 02139 USA Department of Physics and Kavli Institute for Astrophysics and Space Research Massachusetts Institute of Technology 77 Massachusetts Ave Cambridge Massachusetts 02139 USA School of Physics and Astronomy Monash University Vic 3800 Australia OzGrav: The ARC Centre of Excellence for Gravitational Wave Discovery Clayton VIC 3800 Australia

The recent observation of GW190412, the first high-mass ratio binary black hole merger, by the LIGO-Virgo Collaboration (LVC) provides a unique opportunity to probe the impact of subdominant harmonics and precession effects encoded in a gravitational wave signal. We present refined estimates of source parameters for GW190412 using nrsur7dq4, a recently developed numerical relativity waveform surrogate model that includes all ℓ≤4 spin-weighted spherical harmonic modes as well as the full physical effects of precession. We compare our results with two different variants of phenomenological precessing binary black hole waveform models, imrphenompv3hm and imrphenomxphm, as well as to the LVC results. Our results are broadly in agreement with imrphenomxphm results and the reported LVC analysis compiled with the seobnrv4phm waveform model, but in tension with imrphenompv3hm. Using the nrsur7dq4 model, we provide a tighter constraint on the mass ratio (0.26−0.06+0.08) as compared to the LVC estimate of 0.28−0.07+0.13 (both reported as median values with 90% credible intervals). We also constrain the binary to be more face on, and find a broader posterior for the spin precession parameter. We further find that even though ℓ=4 harmonic modes have negligible signal-to-noise ratio, omission of these modes will influence the estimated posterior distribution of several source parameters including chirp mass, effective inspiral spin, luminosity distance, and inclination. We also find that commonly used model approximations, such as neglecting the asymmetric modes (which are generically excited during precession), have negligible impact on parameter recovery for moderate signal-to-noise ratio events similar to GW190412.

关键词： Gravitational waves

Improving early detection of gravitational waves from binary neutron stars using CNNs and FPGAs

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Martins, Ana Lopez, Melissa Baltus, Gregory Meijer, Quirijn van der Sluys, Marc Van Den Broeck, Chris Caudill, Sarah Institute of Theoretical Astrophysics University of Oslo Oslo Norway Utrecht University Netherlands Nikhef – National Institute for Subatomic Physics Netherlands STAR Institut Université de Liège Liège Belgium Department of Physics University of Massachusetts Dartmouth Center for Scientific Computing United States Data Science Research University of Massachusetts Dartmouth Dartmouth United States

The detection of gravitational waves (GWs) from binary neutron stars (BNSs) with possible telescope follow-ups opens a window to ground-breaking discoveries in the field of multi-messenger astronomy. With the improved sensitivity of current and future GW detectors, more BNS detections are expected in the future. Therefore, enhancing low-latency GW search algorithms to achieve rapid speed, high accuracy, and low computational cost is essential. One innovative solution to reduce latency is the use of machine learning (ML) methods embedded in field-programmable gate arrays (FPGAs). In this work, we present a novel WaveNet-based method, leveraging the state-of-the-art ML model, to produce early-warning alerts for BNS systems. Using simulated GW signals embedded in Gaussian noise from the Advanced LIGO and Advanced Virgo detectors’ third observing run (O3) as a proof-of-concept dataset, we demonstrate significant performance improvements. Compared to the current leading ML-based early-warning system, our approach enhances detection accuracy from 66.81% to 76.22% at a 1% false alarm probability. Furthermore, we evaluate the time, energy, and economical cost of our model across CPU, GPU, and FPGA platforms, showcasing its potential for deployment in real-time gravitational wave detection pipelines. Copyright © 2024, The Authors. All rights reserved.

关键词： Alarm systems

Fast And Automatic Floating Point Error Analysis With CHEF-FP

学校读者我要写书评

暂无评论

Fast And Automatic Floating Point Error Analysis With CHEF-F...

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： Garima Singh Baidyanath Kundu Harshitha Menon Alexander Penev David J. Lange Vassil Vassilev European Council for Nuclear Research Meyrin Switzerland Department of Physics Princeton University Princeton New Jersey USA Lawrence Livermore National Laboratory Center for Applied Scientific Computing Livermore California USA Faculty of Mathematics and Informatics University of Plovdiv Plovdiv Bulgaria

As we reach the limit of Moore’s Law, researchers are exploring different paradigms to achieve unprecedented performance. Approximate computing (AC), which relies on the ability of applications to tolerate some error in the results to trade-off accuracy for performance, has shown significant promise. Despite the success of AC in domains such as Machine Learning, its acceptance in High-Performance computing (HPC) is limited due to its stringent requirement of accuracy. We need tools and techniques to identify regions of the code that are amenable to approximations and their impact on the application output quality so as to guide developers to employ selective approximation. To this end, we propose CHEF-FP, a flexible, scalable, and easy-to-use source-code transformation tool based on Automatic Differentiation (AD) for analysing approximation errors in HPC ***-FP uses Clad, an efficient AD tool built as a plugin to the Clang compiler and based on the LLVM compiler infrastructure, as a backend and utilizes its AD abilities to evaluate approximation errors in C++ code. CHEF-FP works at the source level by injecting error estimation code into the generated adjoints. This enables the error-estimation code to undergo compiler optimizations resulting in improved analysis time and reduced memory usage. We also provide theoretical and architectural augmentations to source code transformation-based AD tools to perform FP error analysis. In this paper, we primarily focus on analyzing errors introduced by mixed-precision AC techniques, the most popular approximate technique in HPC. We also show the applicability of our tool in estimating other kinds of errors by evaluating our tool on codes that use approximate functions. Moreover, we demonstrate the speedups achieved by CHEF-FP during analysis time as compared to the existing state-of-the-art tool as a result of its ability to generate and insert approximation error estimate code directly into the derivative sourc

关键词：

A novel signal-consistency test for gravitational-wave searches of generic black hole binaries

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Schmidt, Stefano Caudill, Sarah Nikhef Science Park 105 Amsterdam1098 XG Netherlands Utrecht University Princetonplein 1 Utrecht3584 CC Netherlands Department of Physics University of Massachusetts DartmouthMA02747 United States Center for Scientific Computing and Data Science Research University of Massachusetts DartmouthMA02747 United States

We propose a novel signal-consistency test applicable to a broad search for gravitational waves emitted by generic binary black hole (BBH) systems. The test generalizes the time domain ξ2 signal-consistency test currently utilized by the GstLAL pipeline, which quantifies the discrepancy between the expected signal-to-noise ratio timeseries with the measured one. While the traditional test is restricted to aligned-spin circular orbits and does not account for higher-order modes (HMs), our test does not make any assumption on the nature of the signal. After addressing the mathematical details of the new test, we quantify its advantages in the context of searching for precessing BBHs and/or BBHs with HM content. Our results reveal that for precessing signals, the new test is optimal and has the potential to reduce the values of the ξ2 statistics by up to two orders of magnitude when compared to the standard test. However, in the case of signals with HM content, only a modest enhancement is observed. Recognizing the computational burden associated with the new test, we also derive an approximated signal-consistency test. This approximation maintains the same computational cost as the standard test and can be easily implemented in any matched filtering pipeline with minimal changes, sacrificing only a few percent of accuracy in the low SNR regime. However in the high SNR regime the approximated signal consistency test does not bring any improvement as compared to the "standard" one. By introducing our new test and its approximation and understanding their validity and limitation, this work will benefit any matched-filtering pipeline aimed at searching for BBH signals with strong precession and/or HM content. © 2024, CC BY.

关键词： Pipelines

Learning orbital dynamics of binary black hole systems from gravitational wave measurements

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Keith, Brendan Khadse, Akshay Field, Scott E. Brown University United States Department of Physics and Astronomy The University of Mississippi UniversityMS38677 United States Department of Mathematics Center for Scientific Computing and Visualization Research University of Massachusetts DartmouthMA02747 United States

We introduce a gravitational waveform inversion strategy that discovers mechanical models of binary black hole (BBH) systems. We show that only a single time series of (possibly noisy) waveform data is necessary to construct the equations of motion for a BBH system. Starting with a class of universal differential equations parameterized by feed-forward neural networks, our strategy involves the construction of a space of plausible mechanical models and a physics-informed constrained optimization within that space to minimize the waveform error. We apply our method to various BBH systems including extreme and comparable mass ratio systems in eccentric and non-eccentric orbits. We show that the resulting differential equations apply to time durations longer than the training interval, and relativistic effects, such as perihelion precession, radiation reaction, and orbital plunge, are automatically accounted for. The methods outlined here provide a new, data-driven approach to studying the dynamics of binary black hole systems. Copyright © 2021, The Authors. All rights reserved.

关键词： Constrained optimization