Reliable computing with unreliable components: Using separable environments to stabilize long-term information storage

How, in the face of both intrinsic and extrinsic volatility, can unconventional computing fabrics store information over arbitrarily long periods? Here, we argue that the predictable structure of many realistic environments, both natural and artificial, can be used to maintain useful categorical boundaries even when the computational fabric itself is inherently volatile and the inputs and outputs are partially stochastic. As a concrete example, we consider the storage of binary classifications in connectionist networks, although the underlying principles should be applicable to other unconventional computing paradigms. Specifically, we demonstrate that an unsupervised, activity dependent plasticity rule, AHAH (Anti-Hebbian-And-Hebbian), allows binary classifications to remain stable even when the underlying synaptic weights are subject to random noise. When embedded in environments composed of separable features, the weight vector is restricted by the AHAH rule to local attractors representing stable partitions of the input space, allowing unsupervised recovery of stored binary classifications following random perturbations that leave the system in the same basin of attraction. We conclude that the stability of long-term memories may depend not so much on the reliability of the underlying substrate, but rather on the reproducible structure of the environment itself, suggesting a new paradigm for reliable computing with unreliable components.     

Hamiltonian Forms for a Hierarchy of Discrete Integrable Coupling Systems

A semi-direct sum of two Lie algebras of four-by-four matrices is presented, and a discrete four-by-four matrix spectral problem is introduced. A hierarchy of discrete integrable coupling systems is derived. The obtained integrable coupling systems are all written in their Hamiltonian forms by the discrete variational identity. Finally, we prove that the lattice equations in the obtained integrable coupling systems are all Liouville integrable discrete Hamiltonian systems.     

一种基于图像特征和神经网络的苹果图像分割算法

苹果识别是开发苹果采摘机器人的关键环节,利用图像处理技术和神经网络分类器探索苹果图像分割算法.从苹果树图片中选取苹果图像样本和背景网像样本.分别计算这两类图像样本的颜色特征和纹理特征.颜色特征的计算基于RGB色彩模型,纹理特征的计算基于灰度共生矩阵.选取适当的颜色特征(R/B值)和纹理特征(对比度值和相关性值)作为输入节点,利用反向传播神经网络分类器建模,输出值是一个O～1之间的计算值.通过阈值将输出结果分类为苹果或背景.试验结果表明,该算法正确率大于87.6%,对光照的影响不敏感,是一利较为实用的苹果分割算法.     

Anisotropic Hubbard model on a triangular lattice — spin dynamics in HoMnO3

The recent neutron scattering data for spin-wave dispersion in HoMnO₃ are well-described by an anisotropic Hubbard model on a triangular lattice with a planar (XY) spin anisotropy. Best fit indicates that magnetic excitations in HoMnO₃ correspond to the strong-coupling limit U/t >∼ 15, with planar exchange energy J = 4t ² /U ≃ 2.5 meV and planar anisotropy ΔU ≃ 0.35 meV.      

Probabilistic signatures of spatiotemporal intermittency in the coupled sine circle map lattice

The phase diagram of the coupled sine circle map system exhibits a variety of interesting phenomena including spreading regions with spatiotemporal intermittency, non-spreading regions with spatial intermittency, and coherent structures termed solitons. A spreading to non-spreading transition is seen in the system. A cellular automaton version of the coupled system maps the spreading to non-spreading transition to a transition from a probabilistic to a deterministic cellular automaton. The solitonic sector of the system shows spatiotemporal intermittency with soliton creation, propagation and absorption. A probabilistic cellular automaton mapping is set up for this sector which can identify each one of these phenomena.      

Polarized Raman scattering and lattice eigenmodes of antiferromagnetic NdFeO3

The first‐ and second‐order Raman‐active phonons in the orthorhombic Pbnm NdFeO₃ single crystals were studied by means of polarized Raman scattering and lattice dynamics computations (LDC). The zone‐center phonons of A_g symmetry were distinguished from the B_1g eigenmodes by performing polarized Raman scattering experiments using two parallel polarization configurations, X′(ZZ)X′ and Z(X′X′)Z. With the help of LDC, we were able to assign most of the observed Raman‐active modes, including phonons of B_2g and B_3g symmetry. The LDC results indicated that among the 16 force constants employed, the one corresponding to the stretching vibration between the central Fe cation and the axial oxygen atom in a FeO₆ octahedron unit had the largest value. This suggests that the B‐site Fe cation is more tightly bound to the axial O1 ion than the other two equatorial O2 ions. It was further shown that at higher wavenumbers, the displacement of oxygen atoms contributed dominantly to the zone‐center vibrations. Copyright © 2008 John Wiley & Sons, Ltd.     

Investigation of stimulated dynamics in strongly anisotropic high-temperature superconductors system Bi-Pb-Sr-Ca-Cu-O

It is used the mechanical method of Abrikosov vortex stimulated dynamics investigation in superconductors. With its help it was studied relaxation phenomena in vortex matter of high-temperature superconductors. It established that pulsed magnetic fields change the course of relaxation processes taking place in vortex matter. The study of the influence of magnetic pulses differing by their durations and amplitudes on vortex system of strongly anisotropic high-temperature superconductors system Bi_1.7Pb_0.3Sr₂Ca₂Cu₃O_10−δ showed the presence of threshold phenomena. The small duration pulses do not change the course of relaxation processes taking place in vortex matter. When the duration of pulses exceeds some critical value (threshold), then their influence changes the course of relaxation process which is revealed by stepwise change of relaxing mechanical moment τrel. These investigations showed that the time for formatting of Abrikosov vortex lattice in Bi_1.7Pb_0.3Sr₂Ca₂Cu₃O_10−δ is of the order of 150 μs which on the order of value exceeds the time necessary for formation of a single vortex observed in isotropic high-temperature superconductor HoBa₂Cu₃O_7−δ and on two orders exceeds the creation time of a single vortex observed in classical type II superconductors.     

Raman spectra of crystalline iron perchlorate hexahydrate

Raman studies of crystalline iron perchlorate hexahydrate (Fe(ClO₄)₂·6H₂O) in the region of lattice and anion internal modes were carried out in the temperature range 80–385 K. The temperature‐dependent Raman results are consistent with those from previous works, showing that two phase transitions occur around 336 and 245 K. The transition at 336 K may be considered as an order–disorder transformation, while the one at 245 K is associated with the configurational disorder of the tetrahedron of the perchlorate ions. Copyright © 2008 John Wiley & Sons, Ltd.     

Dimers on the kagome lattice II: Correlations and the Grassmannian approach

In this paper we continue our consideration of closed-packed dimers on the kagome lattice. Using the Pfaffian approach we evaluate the correlation between dimers on two lattice edges. It is found that the correlation is extremely short-ranged in the case of symmetric dimers weights. Explicit expressions for the nonvanishing correlations are obtained for two dimers in the interior of a large lattice. We also describe a Grassmannian functional integral approach, and use it to evaluate the dimer generating function and correlation functions.     

Flow separation behind a rib in a channel with laminar flow

Visualization data and results of combined measurements of flow quantities in flow with separation past a rib at nominally laminar regime of channel flow are reported. In the separation region, the flow is found to be essentially three-dimensional and unsteady, exhibiting a distinct cellular structure and flow zones with transverse motion. It is shown that the rib-induced flow separation gives rise to low-frequency fluctuations of flow velocity and initiates the turbulence transition in the channel flow. The critical Reynolds number at which flow instability starts developing in the channel is estimated. It is shown that at Reynolds numbers higher than the critical Reynolds number the linear integral scale of flow velocity fluctuations in the channel is defined by the duct size.