广播内存网在实时半实物仿真系统中的实现

针对实时半实物仿真系统对实时性要求高的特点,提出一种利用广播内存网技术构建实时半实物仿真系统的方法.详细地阐述了广播内存网的工作原理、技术特点和连接方式,设计了一套基于虚拟共享内存中间件技术的实时半实物仿真基础平台软件.以该基础平台软件为核心,对整个实时半实物仿真系统进行了详尽的功能及性能测试,测试结果表明该系统传输延迟及抖动小,传输可靠性高,可以满足实时半实物仿真系统对实时性和可靠性的要求.     

On Memory Capacity of the Probabilistic Logic Neuron Network

In this paper,the memory capacity of Probabilistic Logic Neuron(PLN) network is discussed.We obtain two main results:(1)the method for constructing a PLN network with a given memory capacity;(2)the relationship between the memory capacity and the size of a PLN network.We show that the memory capacity of a PLN network depends on not only the number of input ports of its element but also the number of elements themselves.The results provide a new method for designing a PLN network.     

Design and Simulation of the Aquarius-II Multiprocessor

Aquarius-II is a cache coherent multiprocessor system designed for the parallel execution of Prolog programs. It contains two tiers of memory: synchronization memory and high bandwidth (HB) memory. The synchronization memory consists of snooping caches connected to a bus and is used to store rendezvous points, synchronization bits, synchronization variables such as locks and semaphores and most of the write shared data. The HB memory is used to store the bulk of the application program code and data. It contains caches and an inexpensive VLSI chip based crossbar interconnection network to memory. The caches connected to the crossbar do not have full snooping capability. The architecture is evaluated by a full simulation of parallel execution of Prolog programs on Aquarius-II. The design details of the components of the architecture and simulation results are presented. Simulation results indicate that the two tier memory system significantly reduces memory interference and speeds up synchronization when compared to a single bus multi. This shared memory multiprocesor architecture has the potential to support other parallel programming paradigms.     

一种基于稀疏分布记忆模型的汉字联想记忆方法

本文描述了Ｋａｎｅｒｖａ的稀疏分布记忆模型，指出了它在用于汉字联想时的问题，同时提出了改进的模型，试验表明，这种改进模型使记忆容量和容错能力大大提高。     

反射内存网实时通信技术的研究

介绍了一种实时网络技术一反射内存网技术，结合实际详细地阐述了反射内存网的工作原理，并与传统计算机网络技术进行了比较，通过设计 一套基于反射内存网的实时通信软件，验证了反射内存网的基本性能。     

语音词汇网单词联想记忆

语音联想记忆是一种高效的记忆方法。为了给学习者提供语音联想的素材,引导学习者进行语音联想,熟悉读音规则,加深对单词拼写和发音的记忆,帮助学习者建立字母组合与相关发音间的双向认知,提出一种基于语音的词汇网。语音词汇网是基于常见字母组合和单词读音之间的差异构建的,因此其中既包含了语音近似度信息,也包含了一定的单词结构信息。利用该网络,学习系统不但可以实现语音联想功能,还能提供语音方面的相关统计信息。语音词汇网的引入能进一步完善单词学习系统的联想记忆功能。     

相位转移熵在再认记忆脑网络中的应用研究

为研究具有不同先验记忆群体的再认记忆脑功能网络差异,引入相位转移熵方法,分析36名中国医学生与37名中国非医学生执行图片再认任务时的脑功能网络。将脑电信号经希尔伯特变换转为瞬时相位数据,使用相空间分箱和实验折叠相结合的方法计算出边际熵和联合熵项,根据相位转移熵定义计算导联之间的相位转移熵,以导联为网络节点,将相位转移熵作为网络的边,并结合复杂网络方法对网络进行分析。结果发现,在20～30 Hz频率范围内,医学生的节点出强度、入强度、局部效率和全局效率（P <0.05）低于非医学生;在额叶部位,医学生的网络枢纽强度均小于非医学生;在左颞叶和右枕叶之间,医学生的大脑信息流向与非医学生相反。相比传统方法,脑网络能够挖掘出更丰富的差异信息。该结果为再认记忆的脑网络研究提供支持。     

基于RapidIO和存储映射的高速互连网络

分析当前高速互连网络中同时存在的TCP/IP, GAMMA, InfiniBand, SCI等技术的实现机制,介绍RapidIO高性能总线技术。研究RapidIO协议和MPC8548处理器的相关技术,提出在RapidIO高速互连网络中实现存储映射的通信技术解决方案。     

概率逻辑神经元网络的记忆容量

本文研究概率逻辑神经元网络的记忆容量问题，主要结果有，１给出构造指定记忆容量网络的方法２给出网络独立记忆容量与网络规模之间的数量关系，得出ＰＬＮ网络的记忆容量不仅与元件的输入线个数有关，而且与网络的元件个数也有关。以上结论为ＰＬＮ网络的综合提供了新方法。     

基于结构学习和迭代自映射的自联想记忆模型

传统的人工神经元网络连接结构是固定的,是对权值的学习.提出一种基于生理神经元特征的人工神经元模型,并在以此为单元构成的用于实现自联想记忆的神经网络上进行对结构的学习.学习算法以设定神经元的输入/输出感受野、调整突触和轴突末梢的连接、并行的自投影迭代为特征.给出了此网络模型的矩阵描述和实验结果.     

LNMS网络内存系统及性能优化技术

在Linux环境下设计并实现了一个两级服务结构的网络内存系统(LNMS)。LNMS较传统网络内存系统具有更好的可扩展性。对LNMS提出了两种性能优化技术：预取和主动内存技术。在预取优化方面,针对并发应用提出一种M-PPM算法。主动内存技术则发掘了内存服务器的计算能力。实验表明,两种优化技术可有效地提升网络内存系统的性能。     

流媒体点播系统关键技术实现

详细介绍系统在实现过程中使用的关键技术,主要包括框架和引擎的设计,使用的平台技术以及独创的一些问题解决方法。由于客户端实现与服务器实现的巨大差异,将分别从客户端和服务器的角度展开讨论。     

Analysis and design of associative memories based on recurrent neural network with discontinuous activation functions

This paper considers a recurrent neural network (RNN) with a special class of discontinuous activation function which is piecewise constants in the state space. One sufficient condition is established to ensure that the novel recurrent neural networks can have (4k−1)ⁿ locally exponential stable equilibrium points. Such RNN is suitable for synthesizing high-capacity associative memories. The design procedure is presented with the method of singular value decomposition. Finally, the validity and performance of the results are illustrated by use of two numerical examples.     

一种新型双向联想记忆神经网络

提出了一种新型双向联想记忆神经网络,此网络将两个相互关联的模式以模式对的形式存储在由N个连接构成的模式环中,记忆容量为22N数量级,完全消除了假模式对、能够全部或部分地回忆出与输入模式对具有最小Hamming距的被记忆的模式对,同时具有较高的记忆效率和可靠性。连接由“连接状态”和“禁止路径”组成,前者直接存储二进制模式对向量的分量,后者用于消除假模式;此神经网络具有正向联想、逆向联想和自联想方式,使得网络能更灵活有效地满足不同的回忆要求。     

混沌神经网络在分离叠加模式和多对多联想记忆中的应用

混沌是不含外加随机因素的完全确定性的系统表现出来的界于规则和随机之间的内秉随机行为。脑神经系统是由神经细胞组成的网络。类似于人脑思维的人工神经网络与冯·诺依曼计算机相比,在信息处理方面有很大的优越性。混沌和神经网络相互融合的研究是从90年代开始的,其主要的目标是通过分析大脑的混沌现象,建立含有混沌动力学的神经网络模型(即混沌神经网络模型),将混沌的遍历性、对初始值敏感等特点与神经网络的非线性、自适应、并行处理优势相结合,     

Space Efficient Quantization for Deep Convolutional Neural Networks

Deep convolutional neural networks (DCNNs) have shown outstanding performance in the fields of computer vision, natural language processing, and complex system analysis. With the improvement of performance with deeper layers, DCNNs incur higher computational complexity and larger storage requirement, making it extremely difficult to deploy DCNNs on resource-limited embedded systems (such as mobile devices or Internet of Things devices). Network quantization efficiently reduces storage space required by DCNNs. However, the performance of DCNNs often drops rapidly as the quantization bit reduces. In this article, we propose a space efficient quantization scheme which uses eight or less bits to represent the original 32-bit weights. We adopt singular value decomposition (SVD) method to decrease the parameter size of fully-connected layers for further compression. Additionally, we propose a weight clipping method based on dynamic boundary to improve the performance when using lower precision. Experimental results demonstrate that our approach can achieve up to approximately 14x compression while preserving almost the same accuracy compared with the full-precision models. The proposed weight clipping method can also significantly improve the performance of DCNNs when lower precision is required.

     

Adapting Memory Hierarchies for Emerging Datacenter Interconnects

E?cient resource utilization requires that emerging datacenter interconnects support both high performance communication and e?cient remote resource sharing. These goals require that the network be mor...     

IPoIB在国产并行系统上的实现与优化

IPoIB是一种在InfiniBand网络上支持IP的协议，使IP应用程序可以运行在InfiniBand网络上.我们在国产并行系统上实现了IPoIB，通过乱序处理、内存拷贝优化、网络参数调优和避免应答延迟的优化手段，实现了IPoIB在国产并行系统上的性能提升.实验结果表明，IPoIB在国产并行系统上正确运行，网络带宽与优化前相比提高近6倍，与10GbE万兆以太网相比，IPoIB更具优势，乱序处理机制减少乱序效果明显.     

直接内存通信技术的研究与实现

为了提高通信网络的带宽、增加通信网络同主机的耦合度,基于共享内存的思想,设计实现了一种直接内存通信网络适配卡(DMC网卡);首先参照内存条的标准,设计DMC网卡与主机之间的接口;使DMC网卡上的存储芯片作为共享内存被CPU和DMC网卡使用;然后,发送数据时,CPU将数据直接写入DMC网卡上的共享内存区;接收时,CPU则直接从共享内存区读取数据,从而减少了通信数据的拷贝次数;最后利用操作系统的内存管理机制,编写驱动程序,实现系统对DMC网卡的管理和操作;经DMC网卡原理样机测试,使用DMC技术能提高网络带宽,改善系统性能;证明了直接内存通信原理的正确性。     

Innovative Hetero-Associative Memory Encoder (HAMTE) for Palmprint Template Protection

Many types of research focus on utilizing Palmprint recognition in user identification and authentication. The Palmprint is one of biometric authentication (something you are) invariable during a person’s life and needs careful protection during enrollment into different biometric authentication systems. Accuracy and irreversibility are critical requirements for securing the Palmprint template during enrollment and verification. This paper proposes an innovative HAMTE neural network model that contains Hetero-Associative Memory for Palmprint template translation and projection using matrix multiplication and dot product multiplication. A HAMTE-Siamese network is constructed, which accepts two Palmprint templates and predicts whether these two templates belong to the same user or different users. The HAMTE is generated for each user during the enrollment phase, which is responsible for generating a secure template for the enrolled user. The proposed network secures the person’s Palmprint template by translating it into an irreversible template (different features space). It can be stored safely in a trusted/untrusted third-party authentication system that protects the original person’s template from being stolen. Experimental results are conducted on the CASIA database, where the proposed network achieved accuracy close to the original accuracy for the unprotected Palmprint templates. The recognition accuracy deviated by around 3%, and the equal error rate (EER) by approximately 0.02 compared to the original data, with appropriate performance (approximately 13 ms) while preserving the irreversibility property of the secure template. Moreover, the brute-force attack has been analyzed under the new Palmprint protection scheme.