3种聚酰亚胺分子的核磁共振碳谱及氢谱的分析

本文报道了３种高性能聚酰亚胺的核磁共振定量碳谱及氢谱。利用这３种样品化学结构的相似性，根据ＰＥＩ和ＤＥＰＴ谱，ＰＥＩ和ＰＯＩ的二维同核化学位移相关谱，碳氢自旋晶格驰豫时间以及化学位移的理论计算，详细分析并归属了３种聚酰亚胺的碳谱峰和氢谱峰，同时验证了其重复单元结构的正确性。     

带干扰风险的破产模型的研究

本文对古典风险模型中保险公司按单位时间常数率收到保险费的假设做了改进,将每次收到的保险费的次数看作是复合泊松过程,将每次收到的保费和每次的理陪额均看作是服从指数分布的随机变量,并引入带干扰风险的扰动项,从而对古典风险模型进行推广,且给出了相应的破产概率上界,分析了破产概率的上界与准备金,索赔额,净保费和扰动方差之间的关系.     

Triplet superconductivity in a 1D itinerant electron system with transverse spin anisotropy

In this paper we study the ground state phase diagram of a one-dimensional t-J-U model away from half-filling. In the large-bandwidth limit and for ferromagnetic exchange with easy-plane anisotropy a phase with gapless charge and massive spin excitations, characterized by the coexistence of triplet superconducting and spin density wave instabilities is realized in the ground state. With increasing ferromagnetic exchange transitions into a ferrometallic and then a spin gapped triplet superconducting phase take place.     

Conductance of nano-systems with interactions coupled via conduction electrons: effect of indirect exchange interactions

A nano-system in which electrons interact and in contact with Fermi leads gives rise to an effective one-body scattering which depends on the presence of other scatterers in the attached leads. This non local effect is a pure many-body effect that one neglects when one takes non interacting models for describing quantum transport. This enhances the non-local character of the quantum conductance by exchange interactions of a type similar to the RKKY-interaction between local magnetic moments. A theoretical study of this effect is given assuming the Hartree-Fock approximation for spinless fermions of Fermi momentum k_F in an infinite chain embedding two scatterers separated by a segment of length L_c. The fermions interact only inside the two scatterers. The dependence of one scatterer onto the other exhibits oscillations of period π/k_F which decay as 1/L_c and which are suppressed when L_c exceeds the thermal length L_T. The analytical results given by the Hartree-Fock approximation are compared with exact numerical results obtained with the embedding method and the DMRG algorithm.     

Quadrupolar susceptibility and magnetic phase diagram of PrNi2Cd20 with non-Kramers doublet ground state

ABSTRACT

In this study, ultrasonic measurements were performed on a single crystal of cubic PrNi₂Cd₂₀, down to a temperature of 0.02?K, to investigate the crystalline electric field ground state and search for possible phase transitions at low temperatures. The elastic constant (C₁₁?C₁₂)/2, which is related to the Γ₃-symmetry quadrupolar response, exhibits the Curie-type softening at temperatures below ～30?K, which indicates that the present system has a Γ₃ non-Kramers doublet ground state. A leveling-off of the elastic response appears below ～0.1?K toward the lowest temperatures, which implies the presence of level splitting owing to a long-range order in a finite-volume fraction associated with Γ₃-symmetry multipoles. A magnetic field–temperature phase diagram of the present compound is constructed up to 28?T for H || [110]. A clear acoustic de Haas–van Alphen signal and a possible magnetic-field-induced phase transition at H ～26?T are also detected by high-magnetic-field measurements.     

Micro-level reserving for general insurance claims using a long short-term memory network

Detailed information about individual claims are completely ignored when insurance claims data are aggregated and structured in development triangles for loss reserving. In the hope of extracting predictive power from the individual claims characteristics, researchers have recently proposed to use micro-level loss reserving approaches. We introduce a discrete-time individual reserving framework incorporating granular information in a deep learning approach named Long Short-Term Memory (LSTM) neural network. At each time period, the network has two tasks: first, classifying whether there is a payment or a recovery, and second, predicting the corresponding non-zero amount, if any. Based on a generalized Pareto model for excess payments over a threshold, we adjust the LSTM reserve prediction to account for extreme payments. We illustrate the estimation procedure on a simulated and a real general insurance dataset. We compare our approach with the chain-ladder aggregate method using the predictive outstanding loss estimates and their actual values.     

Magnetic interactions in strongly correlated systems: Spin and orbital contributions

We present a technique to map an electronic model with local interactions (a generalized multi-orbital Hubbard model) onto an effective model of interacting classical spins, by requiring that the thermodynamic potentials associated to spin rotations in the two systems are equivalent up to second order in the rotation angles, when the electronic system is in a symmetry-broken phase. This allows to determine the parameters of relativistic and non-relativistic magnetic interactions in the effective spin model in terms of equilibrium Green’s functions of the electronic model. The Hamiltonian of the electronic system includes, in addition to the non-relativistic part, relativistic single-particle terms such as the Zeeman coupling to an external magnetic field, spin–orbit coupling, and arbitrary magnetic anisotropies; the orbital degrees of freedom of the electrons are explicitly taken into account. We determine the complete relativistic exchange tensors, accounting for anisotropic exchange, Dzyaloshinskii–Moriya interactions, as well as additional non-diagonal symmetric terms (which may include dipole–dipole interaction). The expressions of all these magnetic interactions are determined in a unified framework, including previously disregarded features such as the vertices of two-particle Green’s functions and non-local self-energies. We do not assume any smallness in spin–orbit coupling, so our treatment is in this sense exact. Finally, we show how to distinguish and address separately the spin, orbital and spin–orbital contributions to magnetism, providing expressions that can be computed within a tight-binding Dynamical Mean Field Theory.     

密度矩阵重正化群的异构并行优化

密度矩阵重正化群方法（DMRG）在求解一维强关联格点模型的基态时可以获得较高的精度,在应用于二维或准二维问题时,要达到类似的精度通常需要较大的计算量与存储空间.本文提出一种新的DMRG异构并行策略,可以同时发挥计算机中央处理器（CPU）和图形处理器（GPU）的计算性能.针对最耗时的哈密顿量对角化部分,实现了数据的分布式存储,并且给出了CPU和GPU之间的负载平衡策略.以费米Hubbard模型为例,测试了异构并行程序在不同DMRG保留状态数下的运行表现,并给出了相应的性能基准.应用于4腿梯子时,观测到了高温超导中常见的电荷密度条纹,此时保留状态数达到10⁴,使用的GPU显存小于12 GB.