Early detection of rotating stall in axial flow compressors via deterministic learning: detectability analysis

Rotating stall and surge are two violent unstable phenomena of an aero-engine compressor. The early detection of rotating stall is a critical and difficult issue in the operation of a compressor. Recently, a deterministic learning based stall inception detection approach (SIDA) has been developed for modeling and detecting stall inception in aero-engine compressors. This paper considers the derivation of analytical results on the detection capabilities for the SIDA based on deterministic learning. First, by utilizing the input/output stability of the residual system, a detectability condition of the SIDA is presented, and how to choose the parameters of the diagnostic system is also analyzed. Second, based on the relationship between NN approximation capabilities and radial basis function (RBF) network structures, the influence of RBF network structures on the performance properties of the SIDA is analyzed. Finally, a simulation study is presented, in which the Mansoux-C2 compressor model is utilized to verify the effectiveness of the proposed SIDA.     

Fusion of learning automata theory and granular inference systems: ANLAGIS. Applications to pattern recognition and machine learning

In this paper the fusion of artificial neural networks, granular computing and learning automata theory is proposed and we present as a final result ANLAGIS, an adaptive neuron-like network based on learning automata and granular inference systems. ANLAGIS can be applied to both pattern recognition and learning control problems. Another interesting contribution of this paper is the distinction between pre-synaptic and post-synaptic learning in artificial neural networks. To illustrate the capabilities of ANLAGIS some experiments on knowledge discovery in data mining and machine learning are presented. The main, novel contribution of ANLAGIS is the incorporation of Learning Automata Theory within its structure; the paper includes also a novel learning scheme for stochastic learning automata.     

Semisupervised Encrypted Traffic Identification Based on Auxiliary Classification Generative Adversarial Network

The rapidly increasing popularity of mobile devices has changed the methods with which people access various network services and increased network traffic markedly. Over the past few decades, network traffic identification has been a research hotspot in the field of network management and security monitoring. However, as more network services use encryption technology, network traffic identification faces many challenges. Although classic machine learning methods can solve many problems that cannot be solved by port- and payload-based methods, manually extract features that are frequently updated is time-consuming and labor-intensive. Deep learning has good automatic feature learning capabilities and is an ideal method for network traffic identification, particularly encrypted traffic identification; Existing recognition methods based on deep learning primarily use supervised learning methods and rely on many labeled samples. However, in real scenarios, labeled samples are often difficult to obtain. This paper adjusts the structure of the auxiliary classification generation adversarial network (ACGAN) so that it can use unlabeled samples for training, and use the wasserstein distance instead of the original cross entropy as the loss function to achieve semisupervised learning. Experimental results show that the identification accuracy of ISCX and USTC data sets using the proposed method yields markedly better performance when the number of labeled samples is small compared to that of convolutional neural network (CNN) based classifier.     

基于DBN深度学习的期货市场价格预测建模与决策

深度学习通过学习深层非线性网络结构即可实现复杂函数的逼近,可以从大量无标注样本集中学习数据集的本质特征。而深度信念网络(DBN)是由多层随机隐变量组成的贝叶斯概率生成模型,可以作为深度神经网络的预训练环节,为该网络提供初始权重。基于该模型的一个高效学习算法不仅解决了模型训练速度慢的问题,还能产生非常好的参数初始值,极大地提升了模型的建模能力。金融市场是一个多变量非线性系统,通过运用DBN模型进行分析预测可以很好地解决其他预测方法初始权重难以确定的问题。文中以原油期货市场价格预测为例,说明了运用DBN模型进行预测和决策的可行性及有效性。     

Deterministic event-based simulation of quantum phenomena

We propose and analyse simple deterministic algorithms that can be used to construct machines that have primitive learning capabilities. We demonstrate that locally connected networks of these machines can be used to perform blind classification on an event-by-event basis, without storing the information of the individual events. We also demonstrate that properly designed networks of these machines exhibit behavior that is usually only attributed to quantum systems. We present networks that simulate quantum interference on an event-by-event basis. In particular we show that by using simple geometry and the learning capabilities of the machines it is possible to simulate single-photon interference in a Mach-Zehnder interferometer. The interference pattern generated by the network of deterministic learning machines is in perfect agreement with the quantum theoretical result for the single-photon Mach-Zehnder interferometer. To illustrate that networks of these machines are indeed capable of simulating quantum interference we simulate, event-by-event, a setup involving two chained Mach-Zehnder interferometers, and demonstrate that also in this case the simulation results agree with quantum theory.     

超闭球CMAC的性能分析及多CMAC结构

如何选择合适网络参数是传统CMAC(Cerebellar Model Articulation Controller)应用 中的一个难题.采用泛化均方差(GMSE)和学习均方差(LMSE)来分别评价超闭球CMAC的泛 化能力与记忆精度,并引入权调整率的概念,来研究CMAC结构参数与学习性能的关系.研究 结果表明,在样本分布和量化级数不变时,泛化均方差和学习均方差是权调整率的非增函数.因 此超闭球CMAC在满足存储空间和计算速度的要求下尽量使得权调整率较大.还提出了并行 CMAC结构以进一步提高单个超闭球CMAC的非线性逼近能力.仿真结果证明了该方法的有 效性.     

用于入侵检测的贝叶斯网络

大型网络的入侵检测主要采用多个分布式代理(Agent)．这些代理具有一定的智能以便处理各种入侵．文章提出用贝叶斯网络构造各Agent，这样的Agent具有学习、快速识别和对不完备数据集的处理能力，从而使系统具有更好的适应性．最后用一实例来说明贝叶斯网络在入侵检测领域内的应用．     

Fault-tolerant training for optimal interpolative nets

The optimal interpolative (OI) classification network is extended to include fault tolerance and make the network more robust to the loss of a neuron. The OI net has the characteristic that the training data are fit with no more neurons than necessary. Fault tolerance further reduces the number of neurons generated during the learning procedure while maintaining the generalization capabilities of the network. The learning algorithm for the fault-tolerant OI net is presented in a recursive formal, allowing for relatively short training times. A simulated fault-tolerant OI net is tested on a navigation satellite selection problem.     

Image segmentation using fuzzy competitive learning based counter propagation network

Image segmentation is the method of partitioning an image into some homogenous regions that are more meaningful for its better understanding and examination. Soft computing methods having the capabilities of achieving artificial intelligence are predominately used to perform the task of segmentation. Due to the variability and the uncertainty present in natural scenes, segmentation is a complicated task to perform with the help of conventional image segmentation techniques. Therefore, in this article a hybrid Fuzzy Competitive Learning based Counter Propagation Network (FCPN) is proposed for the segmentation of natural scene images. This method compromises of the uncertainty handling capabilities of the fuzzy system and proficiency of parallel learning ability of neural network. To identify the number of clusters automatically in less computational time, the instar layer of Counter propagation network (CPN) has been trained by using Fuzzy competitive learning (FCL). The outstar layer of counter propagation network is trained by using Grossberg learning for obtaining the desired output. Region growing method having the tendency to correctly identify edges with simplicity is used for initial seed point selection. Then, the most similar regions in the image are clustered and the number of clusters is estimated automatically. Finally, by identifying the cluster centers the images are segmented. Bacterial foraging algorithm is used to initialize the initial weights to the network, which helps the proposed method in achieving low convergence ratio with higher accuracy. Results validated the higher performance of proposed FCPN method when compared with other states-of-the-art methods. For future work, some other adaptive methods like the fuzzy model-based network can be used to identify multiple object regions and classifying them among separate clusters.

     

GA-based learning for rule identification in fuzzy neural networks

Employing an effective learning process is a critical topic in designing a fuzzy neural network, especially when expert knowledge is not available. This paper presents a genetic algorithm (GA) based learning approach for a specific type of fuzzy neural network. The proposed learning approach consists of three stages. In the first stage the membership functions of both input and output variables are initialized by determining their centers and widths using a self-organizing algorithm. The second stage employs the proposed GA based learning algorithm to identify the fuzzy rules while the final stage tunes the derived structure and parameters using a back-propagation learning algorithm. The capabilities of the proposed GA-based learning approach are evaluated using a well-examined benchmark example and its effectiveness is analyzed by means of a comparative study with other approaches. The usefulness of the proposed GA-based learning approach is also illustrated in a practical case study where it is used to predict the performance of road traffic control actions. Results from the benchmarking exercise and case study effectively demonstrate the ability of the proposed three stages learning approach to identify relevant fuzzy rules from a training data set with a higher prediction accuracy than alternative approaches.     

基于GeoAgent的数字地球自适应网络系统模型

数字地球是一个复杂,动态变化的自适应网络系统。通过类比人类对外界刺激的反应过程和数字地球的行为,提出了以一种具有学习和空间处理能力的智能实体－空间智能体ＧｅｏＡｇｅｎｔ作为数字作自适应网络中的基本处理单元,并给出了基于ＧｅｏＡｇｅｎｔ的数字地球自适应网络的系统模。     

Problem Solving with a Perpetual Evolutionary Learning Architecture

The capability of learning in an indefinite amount of time renders biological systems highly adaptable. We have developed a biologically motivated computer model, called the artificial neuromolecular (ANM) system, that demonstrates long-term evolutionary learning capability for complex problem solving. The major elements of the system are neurons whose input-output behavior is controlled by significant internal dynamics. The dynamics are modeled by cellular automata, structured to represent the neuronal cytoskeleton (a subneuronal network found in every neuron). Neurons of this type are linked into a multilayer network that abstracts some features of visual circuitry. Multiple copies of these networks are controlled by neurons with memory manipulation capabilities. The ANM system combines these two types of neurons into a single, closely integrated architecture. The system is educated to perform desired tasks by evolutionary algorithms. These algorithms act at the intraneuronal level to generate a repertoire of neurons with different pattern processing capabilities. They also act at the interneuronal level (through the memory manipulation system) to orchestrate different pattern processing neurons into a group suitable for performing desired tasks. The system has been applied to Chinese character recognition. Experiments were emphasized on long-term evolutionary learning, relearning capability, self-organizing dynamics, malleability, gradual transformability, multidimensional fitness surface, co-evolutionary learning, and cross-level synergy.     

网络安全态势感知研究综述

网络安全态势感知不同于传统的安全措施,它可以对网络中各种活动的行为进行辨识,从宏观的角度进行意图理解和影响评估,进而提供合理的决策支持,在提高网络的监控能力、应急响应能力及预测网络安全的发展趋势等方面都具有重要的意义。分别对态势感知和网络安全态势感知的定义进行了归纳梳理,对经典的态势感知模型和新发展的网络安全态势感知模型进行了总结与对比;介绍了网络安全态势感知的关键技术,主要分为基于层次化分析、机器学习、免疫系统和博弈论的技术;介绍了近年来网络安全态势感知在因特网、工控网和物联网中的应用;对其未来发展趋势和待解决的问题进行了总结与展望。     

Frequency-based multilayer neural network with on-chip learning andenhanced neuron characteristics

A new digital architecture of the frequency-based multilayer neural network (MNN) with on-chip learning is proposed. As the signal level is expressed by the frequency, the multiplier is replaced by a simple frequency converter, and the neuron unit uses the voting circuit as the nonlinear adder to improve the nonlinear characteristic. In addition, the pulse multiplier is employed to enhance the neuron characteristics. The backpropagation algorithm is modified for the on-chip learning. The proposed MNN architecture is implemented on field programmable gate arrays (FPGA) and the various experiments are conducted to test the performance of the system. The experimental results show that the proposed neuron has a very good nonlinear function owing to the voting circuit. The learning behavior of the MNN with on-chip learning is also tested by experiments, which show that the proposed MNN has good learning and generalization capabilities. Simple and modular structure of the proposed MNN leads to a massive parallel and flexible network architecture, which is well suited for VLSI implementation.     

FPGA Implementation of Neurocomputational Models: Comparison Between Standard Back-Propagation and C-Mantec Constructive Algorithm

Recent advances in FPGA technology have permitted the implementation of neurocomputational models, making them an interesting alternative to standard PCs in order to speed up the computations involved taking advantage of the intrinsic FPGA parallelism. In this work, we analyse and compare the FPGA implementation of two neural network learning algorithms: the standard and well known Back-Propagation algorithm and C-Mantec, a constructive neural network algorithm that generates compact one hidden layer architectures with good predictive capabilities. One of the main differences between both algorithms is the fact that while Back-Propagation needs a predefined architecture, C-Mantec constructs its network while learning the input patterns. Several aspects of the FPGA implementation of both algorithms are analyzed, focusing in features like logic and memory resources needed, transfer function implementation, computation time, etc. The advantages and disadvantages of both methods in relationship to their hardware implementations are discussed.     

A review of online learning in supervised neural networks

Learning in neural networks can broadly be divided into two categories, viz., off-line (or batch) learning and online (or incremental) learning. In this paper, a review of a variety of supervised neural networks with online learning capabilities is presented. Specifically, we focus on articles published in main indexed journals in the past 10 years (2003–2013). We examine a number of key neural network architectures, which include feedforward neural networks, recurrent neural networks, fuzzy neural networks, and other related networks. How the online learning methodologies are incorporated into these networks is exemplified, and how they are applied to solving problems in different domains is highlighted. A summary of the review that covers different network architectures and their applications is presented.     

Learning and relocation capabilities of a CAM-brain machine

Two techniques to enhance the capabilities of a CAM-brain machine are proposed. The first is a learning capability that is realized by providing a “decay register” in each neuron cell. The second is a neural network relocation capability that makes it possible to compact the evolved neural network and make room for further evolution. Both techniques operate in an extrinsic manner and are considered supplementary to the intrinsic evolutionary capability of a CAM-brain. This work was presented, in part, at the Third International Symposium on Artificial Life and Robotics, Oita, Japan, January 19–21, 1998     

分布式深度学习训练网络综述

近年来深度学习在图像、语音、自然语言处理等诸多领域得到广泛应用,但随着人们对深度学习的训练速度和数据处理能力的需求不断提升,传统的基于单机的训练过程愈发难以满足要求,分布式的深度学习训练方法成为持续提升算力的有效途径其中训练过程中节点间网络的通信性能至关重要,直接影响训练性能分析了分布式深度学习中的性能瓶颈,在此基础上...     

A Model for Adaptive Information Retrieval

The paper presents a network model that can be used toproduce conceptual and logical schemas for Information Retrievalapplications. The model has interesting adaptability characteristicsand can be instantiated in various effective ways. The paper alsoreports the results of an experimental investigation into theeffectiveness of implementing associative and adaptive retrieval onthe proposed model by means of Neural Networks. The implementationmakes use of the learning and generalisation capabilities of theBackpropagation learning algorithm to build up and use applicationdomain knowledge in a sub-symbolic form. The knowledge is acquiredfrom examples of queries and relevant documents. Three differentlearning strategies are introduced, their performance is analysed andcompared with the performance of a traditional Information Retrievalsystem.     

An adaptive fuzzy neural network for MIMO system modelapproximation in high-dimensional spaces

An adaptive fuzzy system implemented within the framework of neural network is proposed. The integration of the fuzzy system into a neural network enables the new fuzzy system to have learning and adaptive capabilities. The proposed fuzzy neural network can locate its rules and optimize its membership functions by competitive learning, Kalman filter algorithm and extended Kalman filter algorithms. A key feature of the new architecture is that a high dimensional fuzzy system can be implemented with fewer number of rules than the Takagi-Sugeno fuzzy systems. A number of simulations are presented to demonstrate the performance of the proposed system including modeling nonlinear function, operator's control of chemical plant, stock prices and bioreactor (multioutput dynamical system).