In this work, a novel one-time-programmable memory unit based on a Schottky-type p-GaN diode is proposed. During the programming process, the junction switches from a high-resistance state to a low-resistance state through Schottky junction breakdown, and the state is permanently preserved. The memory unit features a current ratio of more than 10³, a read voltage window of 6 V, a programming time of less than 10^?4 s, a stability of more than 10⁸ read cycles, and a lifetime of far more than 10 years. Besides, the fabrication of the device is fully compatible with commercial Si-based GaN process platforms, which is of great significance for the realization of low-cost read-only memory in all-GaN integration.

提出了一种结合深度学习的空间相位解包裹方法，采用基于改进U-Net网络的编码器-解码器架构，同时加入包含双向长短期记忆网络（BILSTM）的CBiLSTM模块，并且结合注意力机制，避免了典型卷积神经网络学习全局空间依赖关系的固有缺陷的同时增强了深度学习模型对相位解包裹任务中的关键信息的关注能力。通过大量的模拟数据，验证了文中方法在严重噪声(SNR=0)、不连续条件和混叠条件下的鲁棒性，在以上三种情况下，同其他深度学习网络模型进行对比，文中所提出的网络模型的归一化均方根误差（NRMSE）分别为0.75%、1.81%和1.68%；结构相似性指数（SSIM）分别为0.98、0.92和0.94；峰值信噪比(PSNR)分别为40.87、32.56、37.38；同时计算时间显著减少，适合应用到需要快速准确的空间相位解包裹任务中去。通过实际测量数据，验证了文中提出网络模型的可行性。该研究将双向长短期记忆网络（BILSTM）和注意力机制同时引入光学相位解包裹问题中，为解决复杂相位场的解包裹提供了新的思路和方案。

认识并降低噪声对传输粒子保真度的影响, 是量子隐形传态重要的研究方向之一。不同于以前的独立噪声和带记忆的Pauli噪声, 研究了记忆幅值阻尼噪声对保真度的影响, 并给出了一个抵抗该噪声信道的方案。该方案通过让参与者在粒子分发前实施弱测量而粒子接收后实施恢复测量的方式提高保真度。研究结果表明, 记忆幅值阻尼信道的记忆因子强度与保真度大小呈正相关, 并且在部分记忆信道与全记忆信道下弱测量与恢复测量方法也能一定程度上提高传输粒子的保真度。

With the rapid development of machine learning, the demand for high-efficient computing becomes more and more urgent. To break the bottleneck of the traditional Von Neumann architecture, computing-in-memory (CIM) has attracted increasing attention in recent years. In this work, to provide a feasible CIM solution for the large-scale neural networks (NN) requiring continuous weight updating in online training, a flash-based computing-in-memory with high endurance (10⁹ cycles) and ultra-fast programming speed is investigated. On the one hand, the proposed programming scheme of channel hot electron injection (CHEI) and hot hole injection (HHI) demonstrate high linearity, symmetric potentiation, and a depression process, which help to improve the training speed and accuracy. On the other hand, the low-damage programming scheme and memory window (MW) optimizations can suppress cell degradation effectively with improved computing accuracy. Even after 10⁹ cycles, the leakage current (I_off) of cells remains sub-10pA, ensuring the large-scale computing ability of memory. Further characterizations are done on read disturb to demonstrate its robust reliabilities. By processing CIFAR-10 tasks, it is evident that ~90% accuracy can be achieved after 10⁹ cycles in both ResNet50 and VGG16 NN. Our results suggest that flash-based CIM has great potential to overcome the limitations of traditional Von Neumann architectures and enable high-performance NN online training, which pave the way for further development of artificial intelligence (AI) accelerators.

阻变器作为一种基于可逆、非易失、阻态突变的信息存储和处理器件, 是解决传统存储器的内在物理限制和冯·诺依曼架构瓶颈问题的核心电子元器件之一, 受到了广泛关注。卤化物钙钛矿具有快速的载流子迁移特性和优异的光电转换性能, 作为阻变功能层赋予光电阻变存储器优异的阻变性能。因此, 近年来卤化物钙钛矿基阻变器的存储和计算应用研究发展迅速。然而, 目前对于卤化物钙钛矿的光电阻变机理尚未形成统一认识。基于此, 本文分析了卤化物钙钛矿阻变存储器的工作机理, 对比分析了卤化物钙钛矿基光电阻变器导电细丝和能级匹配调控特性, 总结了其各种机理的制约因素, 揭示了导电细丝在光场和电场作用下重复形成和断裂, 以及阻变器中卤化物钙钛矿功能层和其他功能层之间肖特基势垒改变, 主导卤化物钙钛矿光电阻变器的开关比、阈值(Set/Reset)电压和阻变器性能稳定性, 并进一步展望卤化物钙钛矿基光电阻变器在新型人工智能仿生突触、存内运算、机器视觉的应用。