Mixed Mode VLSI Implementation of a Neural Associative Memory

A mixed mode digital/analog special purpose VLSI hardware implementation of an associative memory with neural architecture is presented. The memory concept is based on a matrix architecture with binary storage elements holding the connection weights. To enhance the processing speed analog circuit techniques are applied to implement the algorithm for the association. To keep the memory density as high as possible two design strategies are considered. First, the number of transistors per storage element is kept to a minimum. In this paper a circuit technique that uses a single 6-transistor cell for weight storage and analog signal processing is proposed. Second, the device precision has been chosen to a moderate level to save area as much as possible. Since device mismatch limits the performance of analog circuits, the impact of device precision on the circuit performance is explicitly discussed. It is shown that the device precision limits the number of rows activated in parallel. Since the input vector as well as the output vector are considered to be sparsely coded it is concluded, that even for large matrices the proposed circuit technique is appropriate and ultra large scale integration with a large number of connection weights is feasible.     

A boosted voltage generator for low-voltage DRAMs

This paper proposes a new two-stage two-phase VPP charge pump configured in such a manner that the body effect and the threshold voltage loss are eliminated. The newly proposed circuit is fabricated using 0.18 μm triple-well CMOS process and the measurement result shows that the VPP level tracks 3VDD when VDD is above the threshold voltage.     

Optical Diffractive Memories: Angular Selectivity and Diffraction Efficiency in Dichromated Gelatin

Two thousand images of resolution 512×512 pixels as a regular matrix pattern of 10×10 elements are stored, where each element is angularly multiplexed 20 times in a 25 μm thickness of dichromated gelatin emulsion without cross-talk effect. The surface area of the matrix is 1 cm². We show good concordance of the angular selectivity between the experimental result and theory. The diffraction efficiency of each 20 multiplexed images is measured and has nearly the same value. Examples of reconstructed images for multiple applications are given, for example, storage of 160,000 images on a 3′1/2 floppy disc format, which is about 100 min of black and white film. Application can be made to automobile cartography and storage of X-ray images as well as weather forecast images. Colored diffractive images are also possible and are illustrated.     

Formation of C49-TiSi2 in flash memories: a nucleation controlled phenomenon?

The formation of Ti silicides has been examined in flash memories with 0.25 μm linewidth by X-ray diffraction, scanning electron microscopy, and transmission electron microscopy. It has been observed that, after the first rapid thermal process and the selective metal etch, there is no silicide on the source and on a majority of drain contacts while C49-TiSi₂ is found on the gate. A pre-amorphisation implant increases drastically the formation of C49-TiSi₂ in the drain zone while modifications of annealing conditions have little impact. These results indicate that the formation of C49-TiSi₂ is most likely controlled by nucleation and that this nucleation is sensitive to both the width and the length of the reaction zone. The formation of a Ti rich silicide may play an important role in this nucleation by decreasing the driving force for the formation of C49-TiSi₂. Curiously enough, the formation of C49-TiSi₂ appears thus as a major concern for the salicide process in flash memories.     

Left, right, and inner socles of associative systems

We investigate the basic properties of the different socles that can be considered in not necessarily semiprime associative systems. Among other things, we show that the socle defined as the sum of minimal (or minimal and trivial) inner ideals is always an ideal. When trivial inner ideals are included, this inner socle contains the socles defined in terms of minimal left or right ideals. This revised version was published online in June 2006 with corrections to the Cover Date.     

一个基于相关存储的模式识别方法

应用相关的存储技术实现模式识别可缩减存储大小,但由此带来的饱和问题严重影响系统正常工作,为解决此问题,本文在分析现有系统的基础上给出了基于一系列逻辑操作的识别技术.该技术的实质是用一形式化的方法增加输入模式间的线性无关性.通过论证与实例验证,该方法是有效而实用的.     

环上矩阵的广义Moore-Penrose逆

本文给出带有对合的有1的结合环上一类矩阵的广义Moore-Penrose逆存在的充要条件，而这类矩阵概括了左右主理想整环，单Artin环上所有矩阵。     

EXPONENTIAL STABILITY AND PERIODIC SOLUTION OF HYBRID BIDIRECTIONAL ASSOCIATIVE MEMORY NEURAL NETWORKS WITH DISCRETE DELAYS

In this paper, we study the existence, uniqueness, and the global exponential stability of the periodic solution and equilibrium of hybrid bidirectional associative memory neural networks with discrete delays. By ingeniously importing real parameters di > 0 (i = 1,2, …, n) which can be adjusted, making use of the Lyapunov functional method and some analysis techniques, some new sufficient conditions are established. Our results generalize and improve the related results in [9]. These conditions can be used both to design globally exponentially stable and periodical oscillatory hybrid bidirectional associative neural networks with discrete delays, and to enlarge the area of designing neural networks. Our work has important significance in related theory and its application.     

Hochschild Cohomologies of the Algebra of Smooth Functions

Based on the previously developed technique using the Laplace transformation of distributions with compact support, we calculate the Hochschild cohomologies of the algebra of smooth functions on a finite-dimensional real vector space with coefficients in the adjoint representation.     

TUHF代数上的Lie导子

证明了TUHF代数丁上的Lie导子L形如D l．其中D是T上的结合导子，l是从T到它的中心Z上的线性映射且零化T中的括积．