检索结果-内蒙古大学图书馆

Hebbian errors in learning: An analysis using the Oja model

JOURNAL OF THEORETICAL BIOLOGY 2009年第4期258卷 489-501页

作者： Radulescu, Anca Cox, Kingsley Adams, Paul Univ Colorado Boulder CO 80309 USA SUNY Stony Brook Stony Brook NY 11794 USA Kalypso Inst Stony Brook NY USA

Background: Recent work on long term potentiation in brain slices shows that Hebb's rule is not completely synapse-specific, probably due to intersynapse diffusion of calcium or other factors. We previously Suggested that such errors in Hebbian learning might be analogous to mutations in evolution. Methods and findings: We examine this proposal quantitatively, extending the classical Oja unsupervised model of learning by a single linear neuron to include Hebbian inspecificity. We introduce an error matrix E, which expresses possible crosstalk between updating at different connections. When there is no inspecificity, this gives the classical result of convergence to the first principal component of the input distribution (PC1). We show the modified algorithm converges to the leading eigenvector of the matrix EC, where C is the input covariance matrix. In the most biologically plausible case when there are no intrinsically privileged connections, E has diagonal elements Q and off diagonal elements (1 - Q)/(n - 1), where Q the quality, is expected to decrease with the number of inputs n and with a synaptic parameter b that reflects synapse density, calcium diffusion, etc. We study the dependence of the learning accuracy on b, n and the amount of input activity or correlation (analytically and computationally). We find that accuracy increases (learning becomes gradually less useful) with increases in b, particularly for intermediate (i.e., biologically realistic) correlation strength, although some useful learning always occurs up to the trivial limit Q = 1/n. Conclusions and significance: We discuss the relation of our results to Hebbian unsupervised learning in the brain. When the mechanism lacks specificity, the network fails to learn the expected, and typically most useful, result, especially when the input correlation is weak. Hebbian crosstalk would reflect the very high density of synapses along dendrites, and inevitably degrades learning. (c) 2009 Elsevier Ltd.

关键词： Synaptic plasticity Information transfer Neural computation learning algorithm Dynamical system

来源：评论

学校读者我要写书评

暂无评论

DISTRIBUTED learning OF EQUILIBRIA IN A ROUTING GAME

引用

PARALLEL PROCESSING LETTERS 2009年第2期19卷 189-204页

作者： Barth, Dominique Cohen, Johanne Bournez, Olivier Boussaton, Octave Univ Versailles Lab PRiSM 45 Ave Etals Unis F-78000 Versailles France Ecole Polytech Lab Informat F-91128 Palaiseau France LORIA INRIA CNRS UHP F-54602 Villers Les Nancy France

We focus on the problem of learning equilibria in a particular routing game similar to the Wardrop traffic model. We describe a routing game played by a large number of players and present a distributed learning algorithm that we prove to converge weakly to equilibria for the system. The proof of convergence is based on a differential equation governing the global evolution of the system that is inferred from all the local evolutions of the agents in play. We prove that the differential equation converges with the help of Lyapunov techniques.

关键词： learning algorithm Nash equilibria weak convergellce Lyapunov stability routing game

来源：评论

学校读者我要写书评

暂无评论

Detection and Classification of Incipient Faults in Three-Phase Power Transformer Using DGA Information and Rule-based Machine learning Method

引用

JOURNAL OF CONTROL AUTOMATION AND ELECTRICAL SYSTEMS 2020年第5期31卷 1251-1266页

作者： Katooli, Mohsen Savari Koochaki, Amangaldi Islamic Azad Univ Dept Elect Engn Aliabad Katoul Branch Aliabad Katoul Iran

Three-phase transformers (TPT) play a significant and crucial function in the power networks in order to connect the sub-systems and deliver the electrical energy to final customers. The TPT are one of the most high-priced equipment in modern power networks, and therefore their working condition should be constantly monitored to prevent their breakdown, power outages and huge financial damage. Accordingly, this paper presents a hybrid method for detection and classification of incipient faults in TPT using dissolved gas analysis techniques (DGAT) information and rule-based machine learning method. In the developed method, the most informative and important items of DGAT data out of 14 items selected by association rules mining technique (ARMT) are employed as the input of adaptive neuro-fuzzy inference system (ANFIS). The ARMT is implemented to select the items, which have maximum information and can train the ANFIS more accurately. Furthermore, in order to enhance the accuracy of ANFIS and improve its robustness in different implementations, black widow optimization algorithm is applied for ANFIS training. In order to evaluate the performance of developed method on real issues, two industrial data collections obtained from Iran-Transfo Company chemical laboratory and Damavand power substations are used. The obtained results through MATLAB simulations proved that the developed method has high fault detection and classification accuracy, robust function in different implementations, short run time and simple structure.

关键词： BWO Fuzzy rules Item selection learning algorithm Transformer

来源：评论

学校读者我要写书评

暂无评论

A 2-LEVEL COMMITTEE MACHINE - A REPRESENTATION AND A learning PROCEDURE FOR GENERAL PIECEWISE LINEAR DISCRIMINANT FUNCTIONS

引用

PATTERN RECOGNITION 1981年第3期13卷 269-274页

作者： TAKIYAMA, R Department of Visual Communication Design Kyushu Institute of Design Minami-ku Fukuoka 815 Japan

In order to improve the pattern classification power of the committee machine, a two-level committee machine, which is a committee machine with several lower committees, is proposed and the learning algorithm for it is described. The discriminant function realized by the two-level committee machine can be considered as the general piecewise linear discriminant function which includes Chang's definition. (15) The proposed algorithm is a kind of error-correction procedure, and the learning procedure of the usual committee machine and the perceptron are clearly explained as special cases of the proposed algorithm.

关键词： Pattern recognition Piecewise linear discriminant function Perceptron Committee machine learning algorithm Two-level committee machine

来源：评论

学校读者我要写书评

暂无评论

Fuzzy adaptive learning control network with sigmoid membership function

引用

High Technology Letters 2007年第3期13卷 225-229页

作者：邢杰 Xiao Deyun Department of Automation Tsinghua University Beijing 100084 P.R. China

To get simpler operation in modified fuzzy adaptive learning control network （FALCON） in some engineering application, sigmoid nonlinear function is employed as a substitute of traditional Gaussian membership function. For making the modified FALCON learning more efficient and stable, a simulated annealing （SA） learning coefficient is introduced into learning algorithm. At first, the basic concepts and main advantages of FALCON were briefly reviewed. Subsequently, the topological structure and nodes operation were illustrated; the gradient-descent learning algorithm with SA learning coefficient was derived; and the distinctions between the archetype and the modification were analyzed. Eventually, the significance and worthiness of the modified FALCON were validated by its application to probability prediction of anode effect in aluminium electrolysis cells.

关键词： fuzzy adaptive learning control network （FALCON） topological structure learning algorithm sigmoid function gaussian function simulated annealing （SA）

来源：评论

学校读者我要写书评

暂无评论

Comparison between Extreme learning Machine and Wavelet Neural Networks in Data Classification 9

Comparison between Extreme Learning Machine and Wavelet Neur...

引用

9th International Conference on Machine Vision (ICMV)

作者： Siwar, Yahia Salwa, Said Olfa, Jemai Mourad, Zaied Chokri, Ben Amar Natl Sch Engn Sfax Sfax Tunisia

ISBN: (纸本)9781510611320

Extreme learning Machine is a well known learning algorithm in the field of machine learning. It's about a feedforward neural network with a single-hidden layer. It is an extremely fast learning algorithm with good generalization performance. In this paper, we aim to compare the Extreme learning Machine with wavelet neural networks, which is a very used algorithm. We have used six benchmark data sets to evaluate each technique. These datasets Including Wisconsin Breast Cancer, Glass Identification, Ionosphere, Pima Indians Diabetes, Wine Recognition and Iris Plant. Experimental results have shown that both extreme learning machine and wavelet neural networks have reached good results.

关键词： Extreme learning machine Wavelet neural networks learning algorithm Classification feedforward neural network wavelet

来源：评论

学校读者我要写书评

暂无评论

Dynamic self-organized learning for optimizing the complexity growth of radial basis function neural networks

Dynamic self-organized learning for optimizing the complexit...

引用

IEEE International Conference on Industrial Technology

作者： Arisariyawong, S Charoenseang, S Srinakharinwirot Univ Dept Mech Engn Fac Engn Ongkharak 26120 Nakornayok Thailand

ISBN: (纸本)0780376579

This paper proposes a framework of automatically exploring the optimal size of a radial basis function (RBF) neural network. A dynamic self-organized learning algorithm is presented to adapt the structure of the network. The algorithm generates a new hidden unit based on the steady state error of network and the nearest distance from input data to the center of hidden unit. Furthermore, it also detects and removes any insignificant contributing hidden units. For optimizing the complexity growth of RBF neural network, the growing and pruning are combined during adaptation of RBF neural network structure. The examples of nonlinear dynamical system modeling are presented to illustrate the performance of the proposed algorithm.

关键词： self-organized learning learning algorithm radial basis function networks function estimation convergence

来源：评论

学校读者我要写书评

暂无评论

Dimension Reduction by Manifold learning for Evolutionary learning with Redundant Sensory Inputs

Dimension Reduction by Manifold Learning for Evolutionary Le...

引用

2010 IEEE World Congress on Computational Intelligence

作者： Handa, Hisashi Kawakami, Hiroshi Okayama Univ Grad Sch Nat Sci & Technol Okayama 7008530 Japan Kyoto Univ Grad Sch Informat Kyoto 6068530 Japan

ISBN: (纸本)9781424481262

The optimization of the number and the alignment of sensors is quite important task for designing intelligent agents/robotics. Even though we could use excellent learning algorithms, it will not work well if the alignment of sensors is wrong or the number of sensors is not enough. In addition, if a large number of sensors are available, it will cause the delay of learning. In this paper, we propose the use of Manifold learning for Evolutionary learning with redundant sensory inputs in order to avoid the difficulty of designing the allocation of sensors. The proposed method is composed of two stages: The first stage is to generate a mapping from higher dimensional sensory inputs to lower dimensional space, by using Manifold learning. The second stage is using Evolutionary learning to learn control scheme. The input data for Evolutionary learning is generated by translating sensory inputs into lower dimensional data by using the mapping.

关键词： dimension reduction evolutionary computation evolutionary learning higher dimensional sensory input mapping intelligent robotics intelligent robots learning (artificial intelligence) learning algorithm manifold learning number optimization optimisation redundant sensory input sensor allocation sensor placement sensors sensors alignment sensory input translation

来源：评论

学校读者我要写书评

暂无评论

Control by interconnection of a manipulator arm using reinforcement learning

Control by interconnection of a manipulator arm using reinfo...

引用

IEEE International Symposium on Intelligent Control (ISIC)

作者： Nageshrao, S. P. Lopes, G. A. D. Jeltsema, D. Babuska, R. Delft Univ Technol DCSC Mekelweg 2 NL-2628 CD Delft Netherlands Delft Univ Technol Delft Inst Appl Math NL-2628 CD Delft Netherlands

ISBN: (纸本)9781479977888

Control by interconnection (CbI) is a dynamic output-feedback approach used to control port-Hamiltonian (PH) systems. Here, both the plant and the controller are modelled in PH form, in terms of their own Hamiltonians. However, obtaining an appropriate controller Hamiltonian is generally difficult. In this paper, we address this issue by using reinforcement learning (RL). Additionally due to the semi-supervised optimization nature of the RL algorithms, a performance criterion can be readily included in CbI. We demonstrate the usefulness of the proposed learning algorithm for stabilization of a manipulator arm.

关键词： feedback learning (artificial intelligence) manipulators optimisation stability CbI PH system RL algorithm control by interconnection controller Hamiltonian dynamic output-feedback approach learning algorithm manipulator arm performance criterion port-Hamiltonian system reinforcement learning semi-supervised optimization stabilization Adaptation models Control systems Cost function learning (artificial intelligence) Manipulators Mathematical model Symmetric matrices robotic arm Manipulators learning (artificial intelligence) Performance metrics learning algorithms Symmetric matrix Cost functions control systems Mathematical Model Feedback Adaptation models learning

来源：评论

学校读者我要写书评

暂无评论

Power Entangling and Matching in Cognitive Wireless Mesh Networks by Applying Conjecture Based Multi-agent QQ-learning Approach

Power Entangling and Matching in Cognitive Wireless Mesh Net...

引用

Globecom Workshops

作者： Chen, Xianfu Zhao, Zhifeng Zhang, Honggang Zhejiang Univ York Zhejiang Lab Cognit Radio & Green Commun Hangzhou 310027 Peoples R China

ISBN: (纸本)9781424488650

As the scarce spectrum resource is becoming overcrowded, cognitive wireless mesh networks express great flexibility to improve the spectrum utilization by opportunistically accessing the authorized frequency bands. One of the critical challenges for realizing such networks is how to adaptively match transmit powers and allocate frequency resources among secondary users (SUs) of the licensed frequency bands whilst maintaining the Quality-of-Service (QoS) requirement of the primary users (PUs), even in mutually entangled interference environment. In this paper, we discuss the non-cooperative power allocation matching problem in cognitive wireless mesh networks formed by a number of clusters with the consideration of energy efficiency. Due to the secondary users' selfish and spontaneous features, the problem is modeled as a stochastic learning process. We extend the conventional single-agent Q-learning to a multi-user context, coined as QQ-learning, using the framework of stochastic games. Within the multi-agent QQ-learning processes, a learning SU performs Q-function updates based on the conjecture about the other SUs' behaviors. This learning algorithm provably converges given certain restrictions that arise during learning procedure. Numerical experiments are used to verify the performance of our algorithm and demonstrate its effectiveness of improving the energy efficiency.

关键词： Q-function updates cognitive radio cognitive wireless mesh networks energy efficiency frequency bands learning (artificial intelligence) learning algorithm learning procedure multi-agent systems multiagent QQ-learning approach multiagent QQ-learning proce

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

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

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

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

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：