Modewise Johnson–Lindenstrauss embeddings for nuclear many-body theory

Modewise Johnson–Lindenstrauss embeddings for nuclear many-body theory
复制标题

核多体理论的 Modewise Johnson—Lindenstrauss 嵌入

DOI:
10.1140/epja/s10050-023-00999-5
复制
发表时间:
2023
期刊:
The European Physical Journal A
影响因子:
--
通讯作者:
Iwen, M.
Iwen, M.
中科院分区:
--
文献类型:
--
作者:
Zare, A.;Wirth, R.;Haselby, C. A.;Hergert, H.;Iwen, M.

文献摘要

参考文献

被引文献

相似文献

在目前的工作中,我们启动了一个程序,探讨modewise约翰逊-林登施特劳斯嵌入(JLE)作为一种工具,以减少计算成本和内存需求的(核)多体方法。这些嵌入是高维数据张量到低维子空间的随机投影,这些子空间保留了范数和内积等结构特征。随机嵌入技术的一个吸引人的特点是,它们允许对大张量进行不经意和增量压缩,例如,原子核的哈密顿量或波函数的振幅,分解成小得多的随机草图,仍然允许精确计算基态能量和其他可观测量。特别是,随机JLE技术的遗忘性质使得压缩张量成为可能,而无需事先确切知道稍后可能想要近似的可观测量。这为使用太大而无法存储在内存中的张量打开了大门,例如,当前方法中的未截短的三核子力,或者在大的、对称性不受限制的基础上完整的二加三核子汉密尔顿。这种压缩的哈密顿算子可以被存储并在以后相对容易地使用。作为第一步,我们进行了详细的分析JLE的二阶多体微扰理论(MBPT)的核基态观测值,如能量和半径的修正的影响,注意到这些将是主要的修正在一个良好的行为微扰膨胀,和非常重要的隐式贡献,即使在非微扰方法。广泛的闭壳层核,模型空间和最先进的核相互作用的数值实验证明了所提出的方法的有效性和潜力:我们可以压缩核哈密顿量数百至数千倍,而在基态观测值中仅产生1%或更少的平均相对误差。重要的是,我们表明,JLE捕获的高度结构化的哈密顿张量中包含的相关物理信息,尽管它们的随机特性。除了显著的存储节省,实现的压缩意味着多个数量级的减少计算工作时,压缩的哈密顿量用于高阶MBPT或非微扰多体方法。
In the present work, we initiate a program that explores modewise Johnson–Lindenstrauss embeddings (JLEs) as a tool to reduce the computational cost and memory requirements of (nuclear) many-body methods. These embeddings are randomized projections of high-dimensional data tensors onto low-dimensional subspaces that preserve structural features like norms and inner products. An appealing feature of randomized embedding techniques is that they allow for the oblivious and incremental compression of large tensors, e.g., the nuclear Hamiltonian or wave functions amplitudes, into significantly smaller random sketches that still allow for the accurate calculation of ground-state energies and other observables. In particular, the oblivious nature of randomized JLE techniques makes it possible to compress a tensor without knowing in advance exactly what observables one might want to approximate at a later time. This opens the door for the use of tensors that are much too large to store in memory, e.g., untruncated three-nucleon forces in current approaches, or complete two- plus three-nucleon Hamiltonians in large, symmetry-unrestricted bases. Such compressed Hamiltonians can be stored and used later on with relative ease. As a first step, we perform a detailed analysis of a JLE’s impact on the second-order Many-Body Perturbation Theory (MBPT) corrections for nuclear ground-state observables like the energy and the radius, noting that these will be the dominant corrections in a well-behaved perturbative expansion, and highly important implicit contributions even in nonperturbative approaches. Numerical experiments for a wide range of closed-shell nuclei, model spaces and state-of-the-art nuclear interactions demonstrate the validity and potential of the proposed approach: We can compress nuclear Hamiltonians hundred- to thousandfold while only incurring mean relative errors of 1% or less in ground-state observables. Importantly, we show that JLEs capture the relevant physical information contained in the highly structured Hamiltonian tensor despite their random characteristics. In addition to the significant storage savings, the achieved compressions imply multiple order-of-magnitude reductions in computational effort when the compressed Hamiltonians are used in higher-order MBPT or nonperturbative many-body methods.
粒子数非守恒理论中本征哈密顿量的处理
DOI: --
发表时间: 2009
期刊:
影响因子: --
作者:
H. Hergert;R. Roth
通讯作者: R. Roth
DOI: 10.1063/5.0077770
发表时间: 2022-02-07
影响因子: 4.4
作者:
Hohenstein, Edward G.;Fales, B. Scott;Martinez, Todd J.
通讯作者: Martinez, Todd J.
通过张量分解实现完整的 CCSDT 电子结构模型
DOI: --
发表时间: 2019
期刊: --
影响因子: --
作者:
M. Lesiuk
通讯作者: M. Lesiuk
基于从头算MO量子主方程法的动态第一超极化密度分析方法的发展
DOI: --
发表时间: 2012
期刊:
影响因子: --
作者:
H.Kobayashi;H.Morita;M.Yamauchi;R.Ikeda;H.Kitagawa;Y.Kubota;K.Kato;M.Takata;竹内勝彦;乾智也;安孫子翔・仲程司・藤原尚;Yohsuke Yamamoto;大石徹;Masami Sakamoto;田中裕明;Xin-Dong Jiang;岸亮平
通讯作者: 岸亮平
DOI: 10.1021/acs.jctc.1c00933
发表时间: 2021-12-14
影响因子: 5.5
作者:
Lesiuk, Michal
通讯作者: Lesiuk, Michal