Data augmentation for low-resource languages NMT guided by constrained sampling

Data augmentation (DA) is a ubiquitous approach for several text generation tasks. Intuitively, in the machine translation paradigm, especially in low-resource languages scenario, many DA methods have appeared. The most commonly used methods are building pseudocorpus by randomly sampling, omitting, or replacing some words in the text. However, previous approaches hardly guarantee the quality of augmented data. In this study, we try to augment the corpus by introducing a constrained sampling method. Additionally, we also build the evaluation framework to select higher quality data after augmentation. Namely, we use the discriminator submodel to mitigate syntactic and semantic errors to some extent. Experimental results show that our augmentation method consistently outperforms all the previous state-of-the-art methods on both small and large-scale corpora in eight language pairs from four corpora by 2.38–4.18 bilingual evaluation understudy points.     

High‐Entropy Alloy (HEA)‐Coated Nanolattice Structures and Their Mechanical Properties

Nanolattice structure fabricated by two‐photon lithography (TPL) is a coupling of size‐dependent mechanical properties at micro/nano‐scale with structural geometry responses in wide applications of scalable micro/nano‐manufacturing. In this work, three‐dimensional (3D) polymeric nanolattices are initially fabricated using TPL, then conformably coated with an 80 nm thick high‐entropy alloy (HEA) thin film (CoCrFeNiAl_0.3) via physical vapor deposition (PVD). 3D atomic‐probe tomography (APT) reveals the homogeneous element distribution in the synthesized HEA film deposited on the substrate. Mechanical properties of the obtained composite architectures are investigated via in situ scanning electron microscope (SEM) compression test, as well as finite element method (FEM) at the relevant length scales. The presented HEA‐coated nanolattice encouragingly not only exhibits superior compressive specific strength of ≈0.032 MPa kg^?1 m³ with density well below 1000 kg m^?3, but also shows good compression ductility due to its composite nature. This concept of combining HEA with polymer lattice structures demonstrates the potential of fabricating novel architected metamaterials with tunable mechanical properties.

     

Nano functional neural interfaces

Engineered functional neural interfaces (fNIs) serve as essential abiotic–biotic transducers between an engineered system and the nervous system. They convert external physical stimuli to cellular signals in stimulation mode or read out biological processes in recording mode. Information can be exchanged using electricity, light, magnetic fields, mechanical forces, heat, or chemical signals. fNIs have found applications for studying processes in neural circuits from cell cultures to organs to whole organisms. fNI-facilitated signal transduction schemes, coupled with easily manipulable and observable external physical signals, have attracted considerable attention in recent years. This enticing field is rapidly evolving toward miniaturization and biomimicry to achieve long-term interface stability with great signal transduction efficiency. Not only has a new generation of neuroelectrodes been invented, but the use of advanced fNIs that explore other physical modalities of neuromodulation and recording has begun to increase. This review covers these exciting developments and applications of fNIs that rely on nanoelectrodes, nanotransducers, or bionanotransducers to establish an interface with the nervous system. These nano fNIs are promising in offering a high spatial resolution, high target specificity, and high communication bandwidth by allowing for a high density and count of signal channels with minimum material volume and area to dramatically improve the chronic integration of the fNI to the target neural tissue. Such demanding advances in nano fNIs will greatly facilitate new opportunities not only for studying basic neuroscience but also for diagnosing and treating various neurological diseases.     

砂轮平衡精度的最小二乘测试计算法

提出一种测算砂轮不平衡精度的最小二乘新方法 ,给出基本原理和具体算法 .该方法能有效抑制测量误差的影响 ,大大提高测试精度 ;并能同时测出残余不平衡量的大小和相位 .可用于各类平衡器的精度测定     

加固液晶显示器控制电路设计实例

用FPGA+ARM的方式设计低成本加固液晶显示器控制电路.FPGA负责对比度调整,ARM微控制器负责温度、亮度、通信、键盘扫描的控制.创新之处在于用低成本架构取代老方案,缩短开发周期,并引入LED背光源辅助加热、动态背光源降低功耗的思想.     

Robust adaptive control for greenhouse climate using neural networks

This paper presents a general framework for robust adaptive neural network (NN)‐based feedback linearization controller design for greenhouse climate system. The controller is based on the well‐known feedback linearization, combined with radial basis functions NNs, which allows the feedback linearization technique to be used in an adaptive way. In addition, a robust sliding mode control is incorporated to deal with the bounded disturbances and the approximation errors of NNs. As a result, an inherently nonlinear robust adaptive control law is obtained, which not only provides fast and accurate tracking of varying set‐points, but also guarantees asymptotic tracking even if there are inherent approximation errors. Copyright © 2010 John Wiley & Sons, Ltd.     

B/S模式下条形码票证自定义打印的设计和实现

研究B/S模式下的票务管理信息系统中一维条形码票证的打印输出,提出一种基于ActiveX技术和条码打印机的条形码票证自定义打印方案.通过一个实例,详细阐述了该方案的设计和实现过程.该方案很好地解决了B/S模式下条形码票证的打印.     

一种多级运动目标检测算法

为了提高Javed等提出的运动目标三级检测算法的稳定性,文中对其处理方法做出改进;在像素级处理阶段,采用每一个像素点及其邻域组成的集合作为特征矢量来描述图像,对YUV格式的彩色图像的不同颜色分量分别建立混合高斯模型,得到背景模型后,计算出Sobel边缘检测得到的边缘点的统计分布;在区域级处理阶段,将彩色图像分割与背景建模结合起来,得到具有精确边缘的运动目标,并利用边缘信息消除干扰目标;实验结果表明,即使在前景纹理、颜色比较一致且与背景对比不是很明显的情况下,改进后的方法也能完整地检测出运动目标.     

一种面向大规模P2P环境的成员管理机制

基于P2P的系统都严重依赖于底层节点之间形成稳定的具有某种特征的拓扑结构.成员管理通过为节点维护可以直接访问的邻居视图,能够实现根据上层应用的语义组织节点拓扑.然而大规模P2P环境下节点的动态加入/退出、网络状况的动态波动、节点上资源的动态更新以及庞大的节点规模对成员管理的抗扰动性和可伸缩性提出了更高的要求.许多典型的成员管理技术要么忽略了系统的扰动问题,要么忽略了系统的伸缩问题,因而无法在大规模P2P环境下获得良好的效果.提出了一种新的成员管理机制,综合考虑多种系统资源动态变化的因素,通过抗扰动的成员发现协议以及可伸缩的成员维护方法,保证即使在恶劣环境下仍能够为上层应用维护稳定正确的节点拓扑结构.最后通过实验验证了所提出的成员管理机制在大规模P2P环境下的有效性,通过对实验数据的分析得出的结论为进一步的研究工作奠定了基础.