arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.23934quant-phcond-mat.str-elphysics.comp-ph
收藏

用局部几何信息增强虚时演化

Augmenting Imaginary-Time Evolution with Local Geometric Information

IQM Quantum Computers(IQM量子计算机公司)

Carlos L. Benavides-Riveros, Prachi Sharma, Pietropaolo Frisoni, Fedor Šimkovic

AI总结:

提出增强虚时演化(AITE)框架,利用瞬时能量分布的高阶统计结构确定局部最优方向,实现超线性收敛并在有限虚时内精确消除能量误差,优于标准ITE的渐近指数衰减。

摘要:

虚时演化(ITE)是量子模拟和量子多体物理中基态制备算法的基础。在这些方法中,收敛由瞬时态的能量方差控制,导致流仅渐近地接近基态。我们引入了一个增强虚时演化(AITE)框架,该框架用沿局部最优方向的几何信息下降取代能量景观上的标准梯度流,这些方向通过利用瞬时能量分布的高阶统计结构来识别。由此产生的流在整个演化过程中严格优于标准ITE,并表现出两个性质不同的阶段:超线性收敛阶段,随后是一个灭绝阶段,在该阶段能量误差在有限虚时内精确消失,与标准ITE的渐近指数衰减形成鲜明对比。在AITE的零偏斜极限下恢复标准ITE,这意味着加速自然地扩展到更广泛的ITE算法家族。

英文摘要:

Imaginary-time evolution (ITE) underpins a broad family of algorithms for ground-state preparation in quantum simulation and quantum many-body physics. In these methods, convergence is governed by the energy variance of the instantaneous state, causing the flow to approach the ground state only asymptotically. We introduce an augmented imaginary-time evolution (AITE) framework that replaces the standard gradient flow on the energy landscape with a geometrically informed descent along locally optimal directions, which are identified by exploiting the higher-order statistical structure of the instantaneous energy distribution. The resulting flow strictly outperforms standard ITE throughout the entire evolution and exhibits two qualitatively distinct regimes: a superlinear convergence regime, followed by an extinction regime in which the energy error vanishes exactly at a finite value of the imaginary-time flow parameter, in contrast to the asymptotic exponential decay of ITE. Standard ITE is recovered in the zero-skewness limit of AITE, implying that the acceleration extends naturally across the broader ITE algorithmic family.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑