A Local Learning Algorithm for Dynamic Feedforward and Recurrent Networks

Most known learning algorithms for dynamic neural networks in non-stationary environments need global computations to perform credit assignment. These algorithms either are not local in time or not local in space. Those algorithms which are local in both time and space usually cannot deal sensibly with ‘hidden units’. In contrast, as far as we can judge, learning rules in biological systems with many ‘hidden units’ are local in both space and time. In this paper we propose a parallel on-line learning algorithms which performs local computations only, yet still is designed to deal with hidden units and with units whose past activations are ‘hidden in time’. The approach is inspired by Holland's idea of the bucket brigade for classifier systems, which is transformed to run on a neural network with fixed topology. The result is a feedforward or recurrent ‘neural’ dissipative system which is consuming ‘weight-substance’ and permanently trying to distribute this substance onto its connections in an appropriate way. Simple experiments demonstrating the feasibility of the algorithm are reported.     

Synaptic noise as a means of implementing weight-perturbation learning

Weight-perturbation (WP) algorithms for supervised and/or reinforcement learning offer improved biological plausibility over backpropagation because of their reduced circuitry requirements for realization in neural hardware. This paper explores the hypothesis that biological synaptic noise might serve as the substrate by which weight perturbation is implemented. We explore the basic synaptic noise hypothesis (BSNH), which embodies the weakest assumptions about the underlying neural circuitry required to implement WP algorithms. This paper identifies relevant biological constraints consistent with the BSNH, taxonomizes existing WP algorithms with regard to consistency with those constraints, and proposes a new WP algorithm that is fully consistent with the constraints. By comparing the learning effectiveness of these algorithms via simulation studies, it is found that all of the algorithms can support traditional neural network learning tasks and have similar generalization characteristics, although the results suggest a trade-off between learning efficiency and biological accuracy.     

一种圆环面刀具五坐标端铣加工局部干涉检测算法

对圆环面刀具五坐标端铣加工复杂曲面的特性进行了深入分析.在此基础上,提出了一种快速的局部干涉检测算法,首先计算圆环面刀具圆环面的环心圆到工件曲面之间的最小距离,然后将该最小距离与圆环面刀具的圆角小半径进行比较,以实现对局部干涉的判断,并通过调整刀轴矢量对发生局部干涉的刀位进行处理.算例表明,所提出的干涉检测不仅正确稳定,而且具有很高的效率.     

基于经验小波变换和流形学习的滚动轴承故障诊断研究

针对振动信号的非线性、非平稳性和早期故障特征信号难以提取的特点,提出一种基于经验小波变换(EWT)和流形学习约简的故障特征提取方法。首先利用EWT将振动信号分解成不同特征时间尺度的单分量固有模态函数(IMF),然后从包含故障信息的IMFs中提取滚动轴承的时域统计特征、频域统计特征、AR模型自回归系数和功率谱熵,构造高维特征集;再利用线性局部切空间排列(LLTSA)流形学习算法将构造的高维特征集约简为故障区分度更好的低维特征集;最后利用支持向量机(SVM)对提取特征进行故障识别。实验结果表明该特征提取方法对滚动轴承故障诊断准确率更高。     

Kinetics of the oxidation of metals: Parallel circuit for point defects diffusion

A model is proposed for the oxidation of metals including both bulk crystal lattice diffusion and short-circuit diffusion paths. Assuming local equilibrium between point defects in the bulk and in grain boundaries, we obtain an effective parabolic rate constant k_eff=k_bulk [1+ exp-(H/RT],where H is the enthalpy for the reaction between point defects in the bulk and in short circuits, is the fraction of the short-circuit area, and T is temperature. In the case of the high-temperature oxidation of nickel coated with a thin film of vanadium pentoxide, this model yields to a rate law with a critical oxygen pressure P _c. The nickel vanadium oxide located in the short circuits of the growing oxide NiO is liquid below P_c, leading to a rapid short-circuit diffusion. Above this value, a vanadate precipitates and acts as diffusion blocks for migrating Ni ions.     

神经网络学习算法存在的问题及对策

BP算法是前馈神经网络训l练中应用最多的算法，仉其具有收敛慢和陷入局部极值的严重缺点。为了克服其缺点，近年来人们做了大量的工作。本文从改进的BP算法、智能优化算法、混合学习策略等方面评述了目前应用于前馈神经网络优化的算法，对各种算法的优缺点进行了分析，并给出克服现有算法缺陷的对策。     

Teacher-directed learning: information-theoretic competitive learning in supervised multi-layered networks

In this paper, we propose a new type of efficient learning method called teacher-directed learning. The method can accept training patterns and correlated teachers, and we need not back-propagate errors between targets and outputs into networks. Information flows always from an input layer to an output layer. In addition, connections to be updated are those from an input layer to the first competitive layer. All other connections can take fixed values. Learning is realized as a competitive process by maximizing information on training patterns and correlated teachers. Because information is maximized, information is compressed into networks in simple ways, which enables us to discover salient features in input patterns. We applied this method to the vertical and horizontal lines detection problem, the analysis of US–Japan trade relations and a fairly complex syntactic analysis system. Experimental results confirmed that teacher information in an input layer forces networks to produce correct answers. In addition, because of maximized information in competitive units, easily interpretable internal representations can be obtained.     

混合遗传算法在灰铸铁件缺陷诊断神经网络模型中的应用

基于BP网络的强大缺陷诊断功能 ,构建了灰铸铁件缺陷诊断神经网络模型 ;针对BP算法学习网络权值收敛速度慢 ,易陷入局部最优的缺点 ,采用混合遗传算法加速网络权值的学习。     

Reinforcement learning agents providing advice in complex video games

This article introduces a teacher–student framework for reinforcement learning, synthesising and extending material that appeared in conference proceedings [Torrey, L., & Taylor, M. E. (2013)]. Teaching on a budget: Agents advising agents in reinforcement learning. {Proceedings of the international conference on autonomous agents and multiagent systems}] and in a non-archival workshop paper [Carboni, N., &Taylor, M. E. (2013, May)]. Preliminary results for 1 vs. 1 tactics in StarCraft. {Proceedings of the adaptive and learning agents workshop (at AAMAS-13)}]. In this framework, a teacher agent instructs a student agent by suggesting actions the student should take as it learns. However, the teacher may only give such advice a limited number of times. We present several novel algorithms that teachers can use to budget their advice effectively, and we evaluate them in two complex video games: StarCraft and Pac-Man. Our results show that the same amount of advice, given at different moments, can have different effects on student learning, and that teachers can significantly affect student learning even when students use different learning methods and state representations.     

薄板弯曲分析的16节点流形单元

将数值流形方法应用于薄板弯曲变形的分析,采用标准矩形网格作为数学网格,并提出16节点的矩形薄板流形单元,推导出用于薄板弯曲分析的流形格式和单元矩阵;单元节点的覆盖函数采用C0和C1阶的局部多项式函数形式,单元采用覆盖自由度为基本求解变量,解决了有限元法中采用挠度和转角为求解变量的复杂计算过程;该16节点流形单元应用于正方形薄板弯曲变形的实例分析,结果表明,与有限元分析相比,流形单元的计算精度和收敛性可大幅度提高。     

Influence on the oxidation kinetics of metals by control of the structure of oxide scales

An explanation of the deviation from the parabolic law is the treatment which considers both shortcircuit and lattice diffusion in the oxide scale. In this study we examine how the oxidation kinetics are influenced by changing the structure of the scale of copper oxide in order to confirm the role of short-circuit diffusion in determining the oxidation rate. In addition we explain the oxidation kinetics of copper and nickel by using a model of the scale structure which includes recrystallization and grain growth. Results are as follows: (1) The nucleation and growth behavior of oxide have a direct effect on the structure and in turn the oxidation kinetics due to short-circuit diffusion. (2) A modified treatment is valid in the region where volume diffusion and short-circuit diffusion play an important role in which it is necessary to consider the scale structure such as the grain size distribution and the boundary width. (3) When recrystallization takes place it is necessary to consider the model of a two-layered scale structure which is different in properties and morphology. (4) In this region the rate curves are S-shaped when oxide recrystallization takes place and exhibit a transition from a parabolic to an nth-power relationship (n>2) when grain growth takes place.This research was performed at the University of Tokyo in partial fulfillment of the requirements for the degree of Doctor of Engineering.     

Oxidation of high-chromium Ni-Cr alloys

The oxidation of binary Ni-Cr alloys containing 44 and 50 wt. % Cr has been studied over a range of oxygen partial pressures at temperatures between 800 and 1100°C. The effects of cold work, surface preparation, and distribution of the Cr-rich second phase have been studied. The oxidation behavior is complex and cannot be described by a single model. The oxide grows by short-circuit diffusion as well as bulk transport through Cr ₂ O ₃ scales. The scale-growth mechanism includes extensive metal-oxide separation requiring Cr vapor transport to the scale, compressive stresses within the oxide which result in scale bulging and cracking, and the formation of a second oxide layer which results in voids being incorporated into the scale. Any factor which reduces the oxide grain size, such as cold work, finer distribution of the Cr-rich phase or reduced oxygen pressure, results in an increased oxidation rate of binary alloys because of an increased number of grain-boundary short-circuit diffusion paths.This work is based on a portion of a thesis by G. M. Ecer submitted to the University of Pittsburgh in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Metallurgical and Materials Engineering.Formerly graduate student. Department of Metallurgical and Materials Engineering. University of Pittsburgh.     

碱性焊条焊接短路过渡时间的研究

本试验以各种整流焊机作为焊接电源,用MC—80A 动态分析仪为测试手段,对低氢型焊条手弧焊熔滴短路时间的概率分布进行了探讨。论证了低氢型焊条手弧焊短路时间概率分布属于离散型。短路时间最大的概率值所对应的时间为t_0,则有60%以上的短路时间集中在(t_0±3)ms的区域内。本文着重分析了焊条直径、焊接工艺参数、焊机特性对t_0的影响,初步确定在正常焊接条件下,不同的焊条直径时,t_0的变化范围,这对进一步研究低氢型焊条的飞溅与电源特性的关系,对探讨弧焊电源更为合理的动特性考核方法,有一定参考价值。     

Function Optimization using Connectionist Reinforcement Learning Algorithms

Any non-associative reinforcement learning algorithm can be viewed as a method for performing function optimization through (possibly noise-corrupted) sampling of function values. We describe the results of simulations in which the optima of several deterministic functions studied by Ackley were sought using variants of REINFORCE algorithms. Some of the algorithms used here incorporated additional heuristic features resembling certain aspects of some of the algorithms used in Ackley's studies. Differing levels of performance were achieved by the various algorithms investigated, but a number of them performed at a level comparable to the best found in Ackley's studies on a number of the tasks, in spite of their simplicity. One of these variants, called REINFORCE/MENT, represents a novel but principled approach to reinforcement learning in nontrivial networks which incorporates an entropy maximization strategy. This was found to perform especially well on more hierarchically organized tasks.     

Syntactic Neural Networks

We introduce a new connectionist paradigm which views neural networks as implementations of syntactic pattern recognition algorithms. Thus, learning is seen as a process of grammatical inference and recognition as a process of parsing. Naturally, the possible realizations of this theme are diverse; in this paper we present some initial explorations of the case where the pattern grammar is context-free, inferred (from examples) by a separate procedure, and then mapped onto a connectionist paper. Unlike most neural networks for which structure is pre-defined, the resulting network has as many levels as are necessary and arbitrary connections between levels. Furthermore, by the addition of a delay element, the network becomes capable of dealing with time-varying patterns in a simple and efficient manner. Since grammatical inference algorithms are notoriously expensive computationally, we place an important restriction on the type of context-free grammars which can be inferred. This dramatically reduces complexity. The resulting grammars are called ‘strictly-hierarchical’ and map straightforwardly onto a temporal connectionist parser (TCP) using a relatively small number of neurons. The new paradigm is applicable to a variety of pattern-processing tasks such as speech recognition and character recognition. We concentrate here on hand-written character recognition; performance in other problem domains will be reported in future publications. Results are presented to illustrate the performance of the system with respect to a number of parameters, namely, the inherent variability of the data, the nature of the learning (supervised or unsupervised) and the details of the clustering procedure used to limit the number of non-terminals inferred. In each of these cases (eight in total), we contrast the performance of a stochastic and a non-stochastic TCP. The stochastic TCP does have greater powers of discrimination, but in many cases the results were very similar. If this result holds in practical situations it is important, because the non-stochastic version has a straightforward implementation in silicon.     

基于三角网格多节点覆盖的数值流形方法

以任意三角网格作为数学网格,应用无网格法理论,提出了一种新的权函数构造方法,建立基于任意三角网格的一种多节点流形单元覆盖。针对小挠度变形的薄板弯曲问题,推导出应用于薄板问题的流形格式和单元矩阵。单元节点的覆盖位移采用C0,C1阶的局部多项式形式,积分采用面积坐标的hammer积分。算例表明,与有限元分析相比,该方法的流形单元在计算精度和收敛性上,可得到大幅度提高。     

Minimally Connective,Auto-associative,Neural Networks

Classic barriers to using auto-associative neural networks to model mammalian memory include the unrealistically high synaptic connectivity of fully connected networks, and the relative paucity of information that has been stored in networks with realistic numbers of synapses per neuron and learning rules amenable to physiological implementation. We describe extremely large, auto-associative networks with low synaptic density. The networks have no direct connections between neurons of the same layer. Rather, the neurons of one layer are 'linked' by connections to neurons of some other layer. Patterns of projections of one layer on to another which form projective planes, or other cognate geometries, confer considerable computational power an the network.     

Investigation on particle swarm optimisation for feature selection on high-dimensional data: local search and selection bias

Feature selection is an essential step in classification tasks with a large number of features, such as in gene expression data. Recent research has shown that particle swarm optimisation (PSO) is a promising approach to feature selection. However, it also has potential limitation to get stuck into local optima, especially for gene selection problems with a huge search space. Therefore, we developed a PSO algorithm (PSO-LSRG) with a fast “local search” combined with a gbest resetting mechanism as a way to improve the performance of PSO for feature selection. Furthermore, since many existing PSO-based feature selection approaches on the gene expression data have feature selection bias, i.e. no unseen test data is used, 2 sets of experiments on 10 gene expression datasets were designed: with and without feature selection bias. As compared to standard PSO, PSO with gbest resetting only, and PSO with local search only, PSO-LSRG obtained a substantial dimensionality reduction and a significant improvement on the classification performance in both sets of experiments. PSO-LSRG outperforms the other three algorithms when feature selection bias exists. When there is no feature selection bias, PSO-LSRG selects the smallest number of features in all cases, but the classification performance is slightly worse in a few cases, which may be caused by the overfitting problem. This shows that feature selection bias should be avoided when designing a feature selection algorithm to ensure its generalisation ability on unseen data.     

Swept volumes of toroidal cutters using generating curves

Modelling the volume swept by a tool as it moves from one programmed position to another in three, four and five-axis milling is a challenging task. In this paper, a technique for generating the volume swept by a toroidal cutter in its motion along a given tool path, is presented. This technique is based on identifying “generating curves” along the path and connecting them into a solid model of the swept volume. The swept volume simulations are verified experimentally for three test pieces. These are also compared with the simulation results from the Z-map technique. The results of the experimental verification show the method to be accurate to within 10 μm for the three test pieces. Furthermore, the computation time for the new technique is significantly less than the Z-map method. It is concluded that the proposed method has the potential of allowing fast simulation and verification of multi-axis tool paths.     

Magnetic properties of low Mn-doped NiCuZn nanocrystalline ferrites

Low Mn-doped NiCuZn ferrites with compositions of (Ni_0.6Zn_0.3Cu_0.1)_1−xMn_xFe₂O₄ (where x = 0, 0.01, 0.02 and 0.03) were synthesized directly with sol–gel method. The influence of the Mn²⁺ content (parameter x) and the sintered temperature on the microstructure and the magnetic properties of these ferrites were mainly discussed. With the increasing Mn²⁺ content, saturation magnetization (M_s) of the powder samples decreased. Saturation magnetic flux density (B_s) and remnant magnetic flux density (B_r) of the toroidal specimens decreased with the Mn²⁺ content up to x = 0.02, followed by increasing. B_r and B_s both increased with the increasing sintered temperature, while coercivity (H_c) decreased. The real part of permeability (μ′) of the toroidal specimens increased up to x = 0.01, followed by decreasing. The sintered temperature also affected the resonance frequency and the useable frequency (from which the value of tan δ increased more) obviously. Furthermore, the low-frequency permeability and the secondary maximum (the maximum of the permeability appearing the secondary time) both increased with the sintered temperature from 70 to 120 and 150 to about 600.