通过自动微分实现反磁导设计
Inverse magneto-conductance design by automatic differentiation
AI总结:
针对细线磁导量子干涉图样反映微观结构的逆问题,提出基于自动微分的逆向设计方法,可准确生成缺陷位置并适用于多种复杂图样。
摘要:
细线中的磁导常因传导电子的量子干涉而呈现复杂图样。这些图样反映了线内的微观结构,例如缺陷或电势分布。在本研究中,我们提出一种逆向设计方法,用于自动生成展现所需磁导图样的微观结构。我们通过数值证明,该方法能够准确生成线中的缺陷位置,并可有效应用于各种复杂图样。我们还讨论了用于设计便于实验研究的结构的技术。
英文摘要:
Magneto-conductance in thin wires often exhibits complicated patterns due to the quantum interference of conduction electrons. These patterns reflect microscopic structures in the wires, such as defects or potential distributions. In this study, we propose an inverse design method to automatically generate a microscopic structure that exhibits desired magneto-conductance patterns. We numerically demonstrate that our method accurately generates defect positions in wires and can be effectively applied to various complicated patterns. We also discuss techniques for designing structures that facilitate experimental investigation.
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。