Global stability of Hopfield neural networks

This paper gives a condition for the global stability of a continuous-time hopfield neural network when its activation function maybe not monotonically increasing.

Application of the Conjugate Gradient Optimization Algorithm of Neural Network on Laser Machining

在应用激光技术加工复杂曲面时,通常以采样点集为插值点来建立曲面函数,然后实现曲面上任意坐标点的精确定位。人工神经网络的BP算法能实现函数插值,但计算精度偏低,往往达不到插值精确要求,造成较大的加工误差。提出人工神经网络的共轭梯度最优化插值新算法,并通过实例仿真,证明了这种曲面精确定位方法的可行性,从而为激光加工的三维精确定位提供了一种良好解决方案。这种方法已经应用在实际中。

Three-Dimensional Network Model And Its Parameter Calibration

A material model, whose framework is parallel spring-bundles oriented in 3-D space, is proposed. Based on a discussion of the discrete schemes and optimum discretization of the solid angles, a 3-D network cell consisted of one-dimensional components is developed with its geometrical and physical parameters calibrated. It is proved that the 3-D network model is able to exactly simulate materials with arbitrary Poisson ratio from 0 to 1/2, breaking through the limit that the previous models in the literature are only suitable for materials with Poisson ratio from 0 to 1/3. A simplified model is also proposed to realize high computation accuracy within low computation cost. Examples demonstrate that the 3-D network model has particular superiority in the simulation of short-fiber reinforced composites.

Flows with Inertia in a Three-Dimensional Random Fiber Network

A fiber web is modeled as a three-dimensional random cylindrical fiber network. Nonlinear behavior of fluid flowing through the fiber network is numerically simulated by using the lattice Boltzmann (LB) method. A nonlinear relationship between the friction factor and the modified Reynolds number is clearly observed and analyzed by using the Fochheimer equation, which includes the quadratic term of velocity. We obtain a transition from linear to nonlinear region when the Reynolds numbers are sufficiently high, reflecting the inertial effect of the flows. The simulated permeability of such fiber network has relatively good agreement with the experimental results and finite element simulations.

Comparisons of static, quasi-static and dynamic 3D porous media scale network models for two-phase immiscible flow in porous media

A dynamic 3D pore-scale network model is formulated for investigating the effect of interfacial tension and oil-water viscosity during chemical flooding. The model takes into account both viscous and capillary forces in analyzing the impact of chemical properties on flow behavior or displacement configuration, while the static model with conventional invasion percolation algorithm incorporates the capillary pressure only. From comparisons of simulation results from these models. it indicates that the static pore scale network model can be used successfully when the capillary number is low. With the capillary increases due to the enhancement of water viscosity or decrease of interfacial tension, only the quasi-static and dynamic model can give insight into the displacement mechanisms.

Fluid flow induced calcium response in bone cell network

In our previous work, bone cell networks with controlled spacing and functional intercellular gap junctions had been successfully established by using microcontact printing and self assembled monolayers technologies [Guo, X. E., E. Takai, X. Jiang, Q. Xu, G. M. Whitesides, J. T. Yardley, C. T. Hung, E. M. Chow, T. Hantschel, and K. D. Costa. Mol. Cell. Biomech. 3:95-107, 2006]. The present study investigated the calcium response and the underlying signaling pathways in patterned bone cell networks exposed to a steady fluid flow. The glass slides with cell networks were separated into eight groups for treatment with specific pharmacological agents that inhibit pathways significant in bone cell calcium signaling. The calcium transients of the network were recorded and quantitatively evaluated with a set of network parameters. The results showed that 18 alpha-GA (gap junction blocker), suramin (ATP inhibitor), and thapsigargin (depleting intracellular calcium stores) significantly reduced the occurrence of multiple calcium peaks, which were visually obvious in the untreated group. The number of responsive peaks also decreased slightly yet significantly when either the COX-2/PGE(2) or the NOS/nitric oxide pathway was disrupted. Different from all other groups, cells treated with 18 alpha-GA maintained a high concentration of intracellular calcium following the first peak. In the absence of calcium in the culture medium, the intracellular calcium concentration decreased slowly with fluid flow without any calcium transients observed. These findings have identified important factors in the flow mediated calcium signaling of bone cells within a patterned network.

A Wide Band Strong Acoustic Absorption in a Locally Network Anechoic Coating

Composite materials with interpenetrating network structures usually exhibit unexpected merit due to the cooperative interaction. Locally resonant phononic crystals (LRPC) exhibit excellent sound attenuation performance based on a periodical arrangement of sound wave scatters. Inspired by the interpenetrating network structure and the LRPC concept, we develop a locally network anechoic coating (LNAC) that can achieve a wide band of underwater strong acoustic absorption. The experimental results show that the LNAC possesses an excellent underwater acoustic absorbing capacity in a wide frequency range. Moreover, in order to investigate the impact of the interpenetrating network structure, we fabricate a faultage structure sample and the network is disconnected by hard polyurethane (PU). The experimental comparison between the LNAC and the faultage structure sample shows that the interpenetrating network structure of the LNAC plays an important role in achieving a wide band strong acoustic absorption.

Visual pattern recognition network: Its training algorithm and its optoelectronic architecture

A visual pattern recognition network and its training algorithm are proposed. The network constructed of a one-layer morphology network and a two-layer modified Hamming net. This visual network can implement invariant pattern recognition with respect to image translation and size projection. After supervised learning takes place, the visual network extracts image features and classifies patterns much the same as living beings do. Moreover we set up its optoelectronic architecture for real-time pattern recognition. (C) 1996 Optical Society of America

Optical interconnection for neural networks by use of a self-imaging function

A grating-lens combination unit is developed to form a scaling self-transform function that can self-image on scale. Then an array of many such grating-lens units is used for the optical interconnection of a two-dimensional neural network, and experiments are carried out. We find that our idea is feasible, the optical interconnection system is simple, and optical adjustment is easy. (C) 1998 Optical Society of America.

Optical two-dimensional SW-banyan network: Optical implementation, routing control, and determination of permissible permutations

A 2-D SW-banyan network is introduced by properly folding the 1-D SW-banyan network, and its corresponding optical setup is proposed by means of polarizing beamsplitters and 2-D phase spatial light modulators. Then, based on the characteristics and the proposed optical setup, the control for the routing path between any source-destination pair is given, and the method to determine whether a given permutation is permissible or not is discussed. Because the proposed optical setup consists of only optical polarization elements, it is compact in structure, its corresponding energy loss and crosstalk are low, and its corresponding available number of channels is high. (C) 1996 Society of Photo-Optical Instrumentation Engineers.

Routing scheme, permutation properties, and optical implementation of a four-shuffle-exchange-based Omega network

The routing scheme and some permutation properties of a four-shuffle-exchange-based Omega network are discussed. The corresponding optical setup, which is composed of 2-D phase spatial light modulators and calcite plates, is proposed and demonstrated through mapping the inputs to a 2-D array. Instead of one shuffle-exchange followed by one switching operation as in ordinary Omega networks, in our presented system, the shuffle interconnection embraced in the switches is accomplished simply by varying the switching structure of each stage. For the proposed polarization-optical modules, the system is compact in structure, efficient in performance, and insensitive to the environment. (C) 1997 Society of Photo-Optical Instrumentation Engineers.

Effect of network modifiers on spectroscopic properties of erbium-doped phosphate glasses

The integrated absorption cross section Sigma(abs), I peak emission cross section sigma(cmi), Judd-Ofeld intensity parameters Omega(iota) ( t = 2,4,6), and spontaneous emission probability A(R) of Er3+ ions were determined for Erbium doped alkali and alkaline earth phosphate glasses. It is found the compositional dependence of sigma(emi) 5 almost similar to that of Sigma(abs), which is determined by the sum, of Omega(1) (3 Omega(2) + 10 Omega(4) + 21 Omega(6)). In addition, the compositional dependence of Omega(1) was studied in these glass systems. As a result, compared with. Omega(4) and Omega(6) the Omega(2) has a stronger compositional dependence on the ionic radius and content of modifers. The covalency of Er-O bonds in phosphate glass is weaker than that in silicate glass, germanate glass, aluminate glass, and tellurate glass, since Omega(6) of phosphate glass is relatively large. A(R) is affected by the covalency of the Er3+ ion sites and corresponds to the Omega(6) value.

Structural Integration of Tellurium Oxide into Mixed-Network-Former Glasses: Connectivity Distribution in the System NaPO3–TeO2

FSodium phosphate tellurite glasses in the system (NaPO3)(x)(TeO2)(1-x) were prepared and structurally characterized by thermal analysis, vibrational spectroscopy, X-ray photoelectron spectroscopy (XPS) and a variety of complementary solid-state nuclear magnetic resonance (NMR) techniques. Unlike the situation in other mixed-network-former glasses, the interaction between the two network formers tellurium oxide and phosphorus oxide produces no new structural units, and no sharing of the network modifier Na2O takes place. The glass structure can be regarded as a network of interlinked metaphosphate-type P(2) tetrahedral and TeO4/2 antiprismotic units. The combined interpretation of the O 1s XPS data and the P-31 solid-state NMR spectra presents clear quantitative evidence for a nonstatistical connectivity distribution. Rather the formation of homootomic P-O-P and Te-O-Te linkages is favored over mixed P-O-Te connectivities. As a consequence of this chemical segregation effect, the spatial sodium distribution is not random, as also indicated by a detailed analysis of P-31/No-23 rotational echo double-resonance (REDOR) experiments. ACHTUNGTRENUNG(TeO2)1 x were prepared and structurally characterized by thermal analysis,vibrat ional spectroscopy,X-ray photoelectron spectroscopy (XPS) and a variety of complementary solid-state nuclear magnetic resonance (NMR) techniques. Unlike the situation in other mixed-network-former glasses,the interaction between the two network formers tellurium oxide and phosphorus oxide produces no new structural units,and no sharing of the network modifier Na2O takes place. The glass structure can be regarded as a network of interlinked metaphosphate-type P(2) tetrahedral and TeO4/2 antiprismatic units. The combined interpretation of the O 1s XPS data and the 31P solid-state NMR spectra presents clear quantitative evidence for a nonstatistical connectivity distribution. Rather,the formation of homoatomic P O P and Te O Te linkages is favored over mixed P O Te connectivities. As a consequence of this chemical segregation effect,the spatial sodium distribution is not random,as also indicated by a detailed analysis of 31P/23Na rotational echo double-resonance (REDOR) experiments.