A System Design Methodology for Analog Feed Forward Artificial Neural Networks

Today Feed Forward Neural Networks (FFNs) use paradigms tied to mathematical frameworks more than to actual electronic devices. This fact makes analog neural integrated circuits heavy to design. Here we propose an alternative model that can use the native computational properties of the basic electronic circuits. A practical framework is described to train analog FFNs in compliance with the model. This is especially useful whenever the weight storage elements cannot be re-programmed on the fly at a high rate. To show how the capability of such framework can be applied to neural systems with non conventional architectures two cases are presented. The first one is a neural signal processor named NESP which has sigmoidal neurons and the other is an innovative architecture named N-LESS.     

Field-effect transistor memories based on ferroelectric polymers

Field-effect transistors based on ferroelectrics have attracted intensive interests, because of their non-volatile data retention, rewritability, and non-destructive read-out. In particular, polymeric materials that possess ferroelectric properties are promising for the fabrications of memory devices with high performance, low cost, and large-area manufacturing, by virtue of their good solubility, low-temperature processability, and good chemical stability. In this review, we discuss the material characteristics of ferroelectric polymers, providing an update on the current development of ferroelectric field-effect transistors (Fe-FETs) in non-volatile memory applications.     

Assessment methodology of the lateral migration component in data retention of 3D SONOS memories

A new procedure to assess the impact of lateral charge migration from the overall retention transient is developed. This experimental procedure is designed to be applied on any 3D NAND string, and does not require devices specifically designed; this allows to assess the impact of lateral migration in realistic conditions. The experimental procedure relies on comparing retention loss of a standard sample with a “control sample” not suffering from lateral migration, thanks to a pre-conditioning operation. The pre-conditioning procedure to obtain the control samples is discussed in detail and demonstrated through measurements and simulations.     

Reliability impacts of high-speed 3-bit/cell Schottky barrier nanowire charge-trapping memories

This study experimentally examines the reliability impacts of high-speed 3-bit/cell Schottky barrier nanowire charge-trapping memories. Unique Schottky barrier junctions strongly enhance hot-carrier generation, ensuring high-speed multi-level programming at low gate voltages. However, strong injected gate currents might cause potential retention and endurance concerns when the programming voltage is beyond 9 V. The effective number of deep-level traps is insufficient for capturing injected electrons, such that some electrons occupy shallower states, producing retention degradation after thermal stress. The charge-trapping layers are susceptible to additional trap generation under strong gate currents, leading to considerable threshold-voltage shifts after cycling stress. A compromise of cell characteristics exists between excellent reliability and high-speed programming in 3-bit/cell Schottky barrier nanowire cells. The application of sub-8-V multi-level programming can alleviate the potential reliability generated by strong injected currents, preserving a favorable cycling endurance and thermal retention in 3-bit/cell Schottky barrier nanowire charge-trapping cells.     

Air-stable,non-volatile resistive memory based on hybrid organic/inorganic nanocomposites

A non-volatile memory element based on organic/inorganic nanocomposites is presented. The device can be operated in ambient conditions, showing high retention time and long-term life time. The formation/rupture of metallic filaments in the organic matrix is investigated by HR-XPS and ToF-SIMS analysis, and is demonstrated to be the driving mechanism for the resistive switching.     

双向联想记忆神经网络的一种编码策略

本文提出一种双向联想记忆神经网络的按‘位’加权编码策略，并给出了求取权值递推算法，它将Ｋｏｓｋｏ双向联想记忆网络按海明距离进行模式匹配的原则，修正了按加权海明距离进行了模式匹配，从而可以使得对不满足加续性的所谓“病态结构”的一类样本模式集，同样具有良好的联想能力，对二值图象模式存贮，联想的计算机模拟实验表明，此方法具有优良的性能和实用价值。     

铁电存储技术

简要说明了铁电随机存储器的工作原理及特点，详细阐述了阻碍铁电存储技术发展的技术难点，重点讨论了铁电薄膜材料的疲劳机理，并对铁电存储器的发展作了展望。     

Conformational memories with variable bond angles

Conformational Memories (CM) is a simulated annealing/Monte Carlo method that explores peptide and protein dihedral conformational space completely and efficiently, independent of the original conformation. Here we extend the CM method to include the variation of a randomly chosen bond angle, in addition to the standard variation of two or three randomly chosen dihedral angles, in each Monte Carlo trial of the CM exploratory and biased phases. We test the hypothesis that the inclusion of variable bond angles in CM leads to an improved sampling of conformational space. We compare the results with variable bond angles to CM with no bond angle variation for the following systems: (1) the pentapeptide Met-enkephalin, which is a standard test case for conformational search methods; (2) the proline ring pucker in a 17mer model peptide, (Ala)(8)Pro(Ala)(8); and (3) the conformations of the Ser 7.39 chi(1) in transmembrane helix 7 (TMH7) of the cannabinoid CB1 receptor, a 25-residue system. In each case, analysis of the CM results shows that the inclusion of variable bond angles results in sampling of regions of conformational space that are inaccessible to CM calculations with only variable dihedral angles, and/or a shift in conformational populations from those calculated when variable bond angles are not included. The incorporation of variable bond angles leads to an improved sampling of conformational space without loss of efficiency. Our examples show that this improved sampling leads to better exploration of biologically relevant conformations that have been experimentally validated.     

基于示样储频的自适应瞄频噪声干扰设计

传统瞄频噪声干扰在单一时刻仅能干扰一个雷达工作频点。针对在无瞬时测频引导条件下,传统瞄频噪声干扰对频率捷变、自适应跳频等雷达干扰能力受限的问题,文中提出了一种基于示样脉冲数字储频的自适应瞄频噪声干扰,利用示样脉冲存储截获雷达信号前沿样本,运用窄带噪声对截获的雷达起始频率样本进行时域“乘积冶调制,在不测频条件下,实现收发分时体制干扰机对脉间频率捷变雷达及自适应跳频雷达的自动瞄频干扰,数字仿真结果证明了干扰的有效性。     

Polarization-Resolved Broadband MoS2/Black Phosphorus/MoS2 Optoelectronic Memory with Ultralong Retention Time and Ultrahigh Switching Ratio

The rapidly emerging requirement for device miniaturization and structural flexibility make 2D semiconductors and their van der Waals (vdWs) heterostructures extremely attractive for nonvolatile optoelectronic memory (NOM) applications. Although several concepts for 2D NOM have been demonstrated, multi-heterojunction devices capable of further improving storage performance have received little attention. This work reports a concept for MoS₂/black phosphorus (BP)/MoS₂ multi-heterojunction NOM with artificial trap sites through the BP oxidation, in which the trapped holes at BP/PO_x interface intrigue a persistent photoconductivity that hardly recovers within the experimental time scales (exceeding 10⁴ s). As a result of the interfacial trap-controlled charge injection, the device exhibits excellent photoresponsive memory characteristics, including a record high detectivity of ≈1.2 × 10¹⁶ Jones, a large light-to-dark switching ratio of ≈1.5 × 10⁷_, an ultralow off-state current of ≈1.2 pA, and an outstanding multi-bit storage capacity (11 storage states, 546 nC state^–1). In addition, the middle BP layer in the multi-heterojunction enables broadband spectrum distinction (375–1064 nm), together with a high polarization ratio of 8.4. The obtained results represent the significant step toward the high-density integration of optoelectronic memories with 2D vdWs heterostructures.