arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2003.06118cond-mat.str-elhep-thquant-ph
收藏

量子液体中拓扑序的全息纠缠重整化

Holographic entanglement renormalisation of topological order in a quantum liquid

Anirban Mukherjee, Siddhartha Lal

摘要:

我们提出了一种新的动量空间纠缠重整化群(MERG)方案,用于拓扑有序(T. O.)使用由非局域幺正门组成的幺正量子线路,在正方形晶格(\cite{anirbanmotti,anirbanmott2})上的2D哈伯德模型的基态。在每个MERG步骤中,幺正量子线路解开一组电子态,从而变换多粒子态的张量网络表示。通过将非局域么正门表示为两个量子比特解纠缠门的乘积,我们给出了MERG的纠缠全息映射(EHM)表示。利用量子信息理论和复杂网络理论中基于纠缠的测度,我们研究了EHM体中拓扑序的出现。MERG揭示了在文献\cite{anirbanmotti}中发现的半填充时,2D哈伯德模型的正常金属、拓扑有序的绝缘量子液体和Neél反铁磁非对称破缺基态的独特的全息纠缠特征。在参考文献\cite{anirbanmott2}中发现的空穴掺杂的2D哈伯德模型的量子临界点的MERG分析揭示了具有空穴掺杂的量子液体基态的多粒子纠缠的演化,以及Mottness的崩溃如何导致d波超导性的出现。我们执行的EHM网络的信息理论分析,表明信息瓶颈的原则是负责的EHM网络的分层结构中的纠缠功能的蒸馏。因此,我们基于EHM网络构建了一个深度神经网络(DNN)架构,并将其用于预测拓扑顺序的开始。我们还证明了DNN能够区分拓扑有序和无间隙的正常金属相。

英文摘要:

We introduce a novel momentum space entanglement renormalization group (MERG) scheme for the topologically ordered (T.O.) ground state of the 2D Hubbard model on a square lattice (\cite{anirbanmotti,anirbanmott2}) using a unitary quantum circuit comprised of non-local unitary gates. At each MERG step, the unitary quantum circuit disentangles a set of electronic states, thereby transforming the tensor network representation of the many-particle state. By representing the non-local unitary gate as a product of two-qubit disentangler gates, we provide an entanglement holographic mapping (EHM) representation for MERG. Using entanglement based measures from quantum information theory and complex network theory, we study the emergence of topological order in the bulk of the EHM. The MERG reveals distinct holographic entanglement features for the normal metallic, topologically ordered insulating quantum liquid and Neél antiferromagnetic symmetry-broken ground states of the 2D Hubbard model at half-filling found in Ref.\cite{anirbanmotti}. An MERG analysis of the quantum critical point of the hole-doped 2D Hubbard model found in Ref.\cite{anirbanmott2} reveals the evolution of the many-particle entanglement of the quantum liquid ground state with hole-doping, as well as how the collapse of Mottness is responsible for the emergence of d-wave superconductivity. We perform an information theoretic analysis of the EHM network, demonstrating that the information bottleneck principle is responsible for the distillation of entanglement features in the heirarchical structure of the EHM network. As a result, we construct a deep neural network (DNN) architecture based on our EHM network, and employ it for predicting the onset of topological order. We also demonstrate that the DNN is capable of distinguishing between the topologically ordered and gapless normal metallic phases.

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。

↑