检索结果-内蒙古大学图书馆

Evolutionary training of hardware realizable multilayer perceptrons

NEURAL COMPUTING & APPLICATIONS 2006年第1期15卷 33-40页

作者： Plagianakos, VP Magoulas, GD Vrahatis, MN Univ Patras Articicial Intelligence Res Ctr Computat Intelligence Lab Dept Math GR-26110 Patras Greece Univ London Sch Comp Sci & Informat Syst London WC1E 7HX England

The use of multilayer perceptrons (MLP) with threshold functions (binary step function activations) greatly reduces the complexity of the hardware implementation of neural networks, provides tolerance to noise and improves the interpretation of the internal representations. In certain case, such as in learning stationary tasks, it may be sufficient to find appropriate weights for an MLP with threshold activation functions by software simulation and, then, transfer the weight values to the hardware implementation. Efficient training of these networks is a subject of considerable ongoing research. Methods available in the literature mainly focus on two-state (threshold) nodes and try to train the networks by approximating the gradient of the error function and modifying appropriately the gradient descent, or by progressively altering the shape of the activation functions. In this paper, we propose an evolution-motivated approach, which is eminently suitable for networks with threshold functions and compare its performance with four other methods. The proposed evolutionary strategy does not need gradient related information, it is applicable to a situation where threshold activations are used from the beginning of the training, as in "on-chip" training, and is able to train networks with integer weights.

关键词： feedforward neural networks backpropagation algorithm neural networks with threshold activations integer weight neural networks integer programming steepest descent unconstrained optimization differential evolution

来源：评论

学校读者我要写书评

暂无评论

Estimation of chemical oxygen demand by ultraviolet spectroscopic profiling and artificial neural networks

引用

NEURAL COMPUTING & APPLICATIONS 2006年第3-4期15卷 197-203页

作者： Fogelman, S Blumenstein, M Zhao, HJ Griffith Univ Ctr Aquat Proc & Pollut Sch Environm & Appl Sci Fac Environm SciGold Coast Mail Ctr Gold Coast Qld 9726 Australia Griffith Univ Sch Informat & Commun Technol Fac Engn & Informat Technol Gold Coast Mail Ctr Gold Coast Qld 9726 Australia

A simple method based on the mathematical treatment of spectral absorbance profiles in conjunction with artificial neural networks (ANNs) is demonstrated for rapidly estimating chemical oxygen demand (COD) values of wastewater samples. In order to improve spectroscopic analysis and ANN training time as well as to reduce the storage space of the trained ANN algorithm, it is necessary to decrease the ANN input vector size by extracting unique characteristics from the raw input pattern. Key features from the spectral absorbance pattern were therefore selected to obtain the spectral absorbance profile, reducing the ANN input vector from 160 to 10 selected inputs. The results indicate that the COD values obtained from the selected absorbance profiles agreed well with those obtained from the entire absorbance pattern. The spectral absorbance profile technique was also compared to COD values estimated by a multiple linear regression (MLR) model to validate whether ANNs were better and more robust models for rapid COD analysis. It was found that the ANN model predicted COD values closer to standard COD values than the MLR model.

关键词： chemical oxygen demand UV-vis spectroscopy artificial neural networks backpropagation algorithm multiple linear regression

来源：评论

学校读者我要写书评

暂无评论

A NEW-TYPE OF FUZZY NEURAL-NETWORK-BASED ON A TRUTH SPACE APPROACH FOR AUTOMATIC ACQUISITION OF FUZZY RULES WITH LINGUISTIC HEDGES

引用

INTERNATIONAL JOURNAL OF APPROXIMATE REASONING 1995年第4期13卷 249-268页

作者： HORIKAWA, S FURUHASHI, T UCHIKAWA, Y Department of Information Electronics Faculty of Engineering Nagoya University Furo-cho Chikusa-ku Nagoya Japan

Fuzzy reasoning methods are generally classified into two approaches: the direct approach and the truth space approach. Several researches on the relationships between these approaches have been reported. There has been, however;no research which discusses their utility. The authors have previously proposed four types of fuzzy neural networks (FNNs) called Type I, II, III, and IV. The FNNs can identify the fuzzy rules and tune the membership functions of fuzzy reasoning automatically, utilizing the learning capability of neural networks. Types III and IV;which ara based on the truth space approach, can acquire linguistic fuzzy rules with the fuzzy variables in the consequences labeled according to their linguistic truth values (LTVs). However, the expressions available for the linguistic labeling are limited since the LTVs are singletons. This paper presents a new type of FNN based on the truth space approach for automatic acquisition of the fuzzy rules with linguistic hedges. The new FNN, called Type V has the LTVs defined by fuzzy sets for fuzzy rules and can express the identified fuzzy rules linguistically using the fuzzy variables in the consequences with linguistic hedges. Two simulations are done for demonstrating the feasibility of the new method. The results show that the truth space approach makes the fuzzy rules easy to understand.

关键词： FUZZY NEURAL NETWORK TRUTH SPACE APPROACH AUTOMATIC ACQUISITION LINGUISTIC HEDGES backpropagation algorithm FUZZY MODELING FUZZY CONTROL

来源：评论

学校读者我要写书评

暂无评论

IMPROVING NEURAL NET CONVERGENCE

引用

PATTERN RECOGNITION LETTERS 1992年第5期13卷 331-338页

作者： ALDER, M LIM, SG HADINGHAM, P ATTIKIOUZEL, Y UNIV WESTERN AUSTRALIA DEPT ELECTR & ELECT ENGNNEDLANDSWA 6009AUSTRALIA UNIV WESTERN AUSTRALIA DEPT COMP SCINEDLANDSWA 6009AUSTRALIA

This paper shows a certain equivalence between the 3-layer feed forward back-propagation neural net of Rumelhart et al. and the committee net of Nilsson. This is used to produce an improvement in the performance of the former. It is found that (a) the number of epochs taken is reduced by a factor of between 6 and 10, (b) the time taken is reduced by a factor of about 20, and (c) the net converges under conditions when the back-propagation algorithm is trapped in local minima and fails to converge.

关键词： COMMITTEE NET backpropagation algorithm

来源：评论

学校读者我要写书评

暂无评论

Editorial: From Explainable Artificial Intelligence (xAI) to Understandable Artificial Intelligence (uAI)

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第9期5卷 4310-4314页

作者： Abbass, Hussein Crockett, Keeley Garibaldi, Jonathan Gegov, Alexander Kaymak, Uzay Sousa, Joao Miguel C. University of New South Wales School of Systems and Computing CanberraACT2600 Australia Manchester Metropolitan University Department of Computing and Mathematics ManchesterM1 5GD United Kingdom University of Nottingham Ningbo China Ningbo315100 China University of Portsmouth PortsmouthPO1 3HE United Kingdom Eindhoven University of Technology Jheronimus Academy of Data Science s-Hertogenbosch5211 Netherlands IDMEC Instituto Superior Técnico Lisboa1049-001 Portugal

来源：评论

学校读者我要写书评

暂无评论

Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and backpropagation

引用

FRONTIERS IN COMPUTATIONAL NEUROSCIENCE 2017年 11卷 24页

作者： Scellier, Benjamin Bengio, Yoshua Univ Montreal Montreal Inst Learning Algorithms Dept Informat & Rech Operat Montreal PQ Canada

We introduce Equilibrium Propagation, a learning framework for energy-based models. It involves only one kind of neural computation, performed in both the first phase (when the prediction is made) and the second phase of training (after the target or prediction error is revealed). Although this algorithm computes the gradient of an objective function just like backpropagation, it does not need a special computation or circuit for the second phase, where errors are implicitly propagated. Equilibrium Propagation shares similarities with Contrastive Hebbian Learning and Contrastive Divergence while solving the theoretical issues of both algorithms: our algorithm computes the gradient of a well-defined objective function. Because the objective function is defined in terms of local perturbations, the second phase of Equilibrium Propagation corresponds to only nudging the prediction (fixed point or stationary distribution) toward a configuration that reduces prediction error. In the case of a recurrent multi-layer supervised network, the output units are slightly nudged toward their target in the second phase, and the perturbation introduced at the output layer propagates backward in the hidden layers. We show that the signal "back-propagated" during this second phase corresponds to the propagation of error derivatives and encodes the gradient of the objective function, when the synaptic update corresponds to a standard form of spike-timing dependent plasticity. This workmakes it more plausible that a mechanism similar to backpropagation could be implemented by brains, since leaky integrator neural computation performs both inference and error back-propagation in our model. The only local difference between the two phases is whether synaptic changes are allowed or not. We also show experimentally that multi-layer recurrently connected networks with 1, 2, and 3 hidden layers can be trained by Equilibrium Propagation on the permutation-invariant MNIST task.

关键词： artificial neural network backpropagation algorithm biologically plausible learning rule contrastive hebbian learning deep learning fixed point Hopfield networks spike-timing dependent plasticity

来源：评论

学校读者我要写书评

暂无评论

Applications of artificial neural networks for patterning and predicting aquatic insect species richness in running waters

引用

ECOLOGICAL MODELLING 2003年第3期160卷 265-280页

作者： Park, YS Céréghino, R Compin, A Lek, S Univ Toulouse 3 CNRS UMR 5576 CESAC F-31062 Toulouse France

Two artificial neural networks (ANNs), unsupervised and supervised learning algorithms, were applied to suggest practical approaches for the analysis of ecological data. Four major aquatic insect orders (Ephemeroptera, Plecoptera, Trichoptera, and Coleoptera, i.e. EPTC), and four environmental variables (elevation, stream order, distance from the source, and water temperature) were used to implement the models. The data were collected and measured at 155 sampling sites on streams of the Adour-Garonne drainage basin (South-western France). The modelling procedure was carried out following two steps. First, a self-organizing map (SOM), an unsupervised ANN, was applied to classify sampling sites using EPTC richness. Second, a backpropagation algorithm (BP),. a supervised ANN, was applied to predict EPTC richness using a set of four environmental variables. The trained SOM classified sampling sites according to a gradient of EPTC richness, and the groups obtained corresponded,to geographic regions of the drainage basin and characteristics of their environmental variables. The SOM showed its convenience to analyze relationships among sampling sites, biological attributes, and environmental variables. After accounting for the relationships in data sets, the BP used to predict the EPTC richness with a, set of four environmental variables, showed a high accuracy (r = 0.91 and r = 0.61 for training and test data sets respectively). The prediction of EPTC richness is thus a valuable tool to. assess disturbances in given areas: by knowing what the EPTC richness should be, we can determine the degree to which disturbances have altered it. The results suggested that methodologies successively using two different neural networks are helpful to understand ecological data through ordination first, and then to predict target variables: (C) 2002 Elsevier Science B.V. All rights reserved.

关键词： self-organizing-map backpropagation algorithm ordination classification prediction aquatic insects species richness

来源：评论

学校读者我要写书评

暂无评论

NEW TRENDS IN STRUCTURE-BIODEGRADABILITY RELATIONSHIPS

引用

QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS 1993年第1期12卷 49-56页

作者： CAMBON, B DEVILLERS, J CTR TRAITEMENT INFORMAT SCI 21 RUE BANNIEREF-69003 LYONFRANCE

Problems related to structure-biodegradation models are discussed. They deal with the homogeneity of the data sets, the selection of an adequate statistical method, and the choice of the molecular descriptors. This short review shows that the Boolean backpropagation neural networks are promising tools to model biodegradation. To confirm this hypothesis, a heterogeneous learning set of 47 molecules and two testing sets of 23 and 17 chemicals weakly (0) or highly (1) biodegradable are described by means of 11 Boolean structural descriptors. A supervised neural network using the backpropagation algorithm generates 100% and 85% of good predictions for the learning and testing sets, *** results obtained are compared with those obtained from classical regression analysis. Advantages of this new approach are given. A particular application of correspondence factor analysis is presented to transform the Boolean variables before their introduction in the neural network. Advantages of this statistical transformation are stressed.

关键词： backpropagation algorithm BIODEGRADATION BOOLEAN DESCRIPTORS CORRESPONDENCE FACTOR ANALYSIS NEURAL NETWORKS STRUCTURE-BIODEGRADABILITY RELATIONSHIPS

来源：评论

学校读者我要写书评

暂无评论

Design of backpropagation Neural Network for Aging Estimation of Electric Battery

引用

SENSORS AND MATERIALS 2023年第4期35卷 1385-1395页

作者： Shin, Kyoo Jae Busan Univ Foreign Studies Dept Artificial Intelligence Convergence 65 Geumsaem ro 485 Beon gil Busan 46234 South Korea

The state of charge (SOC) of an electric vehicle is very important for predicting the remaining battery level and safely protecting the battery from over-discharge and overcharge conditions. In this regard, a neural network (NN) algorithm using backpropagation (BP) has been proposed to accurately estimate the SOC of a battery. Lithium polymer batteries have a nonlinear relationship between their estimated SOC and the current, voltage, and temperature. In this study, a lithium polymer battery with a capacity of 3.7 V/16 Ah was applied. A charge/discharge experiment was performed under constant current and temperature conditions at a discharge rate of 0.5 C. The experimental data were used to train a backpropagation neural network (BPNN) that was used to predict the SOC under charging conditions and the depth of dispatch (DOD) performance under discharge conditions. As a result of the experiment, the error of the proposed BPNN model was found to be 0.22% of the mean absolute error in the discharge DOD and 0.19% of the mean absolute error in the charging SOC at 10, 50, 100, and 150 cycles. Therefore, the high performance of the SOC learning model of the designed BP algorithm was confirmed.

关键词： electric vehicle battery state of charge machine learning methods neural network backpropagation algorithm

来源：评论

学校读者我要写书评

暂无评论

COMPARISON OF CONVERGENCE BEHAVIOR AND GENERALIZATION ABILITY IN backpropagation LEARNING WITH LINEAR AND SIGMOID OUTPUT UNITS

引用

IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES 1993年第6期E76A卷 1035-1042页

作者： KAMRUZZAMAN, J KUMAGAI, Y HIKITA, H Bangladesh Univ of Engineering and Technology Dhaka Bangladesh

The most commonly used activation function in backpropagation learning is sigmoidal while linear function is also sometimes used at the output layer with the view that choice between these activation functions does not make considerable differences in network's performance. In this letter, we show distinct performance between a network with linear output units and a similar network with sigmoid output units in terms of convergence behavior and generalization ability.(10) We experimented with two types of cost functions, namely, sum-squared error used in standard backpropagation and log-likelihood recently reported.(3),(4) We find that, with sum-squared error cost function and hidden units with nonsteep sigmoid function, use of linear units at the output layer instead of sigmoidal ones accelerates the convergence speed considerably while generalization ability is slightly degraded. Network with sigmoid output units trained by log-likelihood cost function yields even faster convergence and better generalization but does not converge at all with linear output units. It is also shown that a network with linear output units needs more hidden units for convergence.

关键词： backpropagation algorithm ACTIVATION FUNCTION LOG-LIKELIHOOD COST FUNCTION CONVERGENCE BEHAVIOR GENERALIZATION ABILITY

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：