A robust subspace algorithm for principal component analysis

We present a noise robust PCA algorithm which is an extension of the Oja subspace algorithm and allows tuning the noise sensitivity. We derive a loss function which is minimized by this algorithm and interpret it in a noisy PCA setting. Results on the local stability analysis of this algorithm are given and it is shown that the locally stable equilibria are those which minimize the loss function.     

A neural mass model based on single cell dynamics to model pathophysiology

Neural mass models are successful in modeling brain rhythms as observed in macroscopic measurements such as the electroencephalogram (EEG). While the synaptic current is explicitly modeled in current models, the single cell electrophysiology is not taken into account. To allow for investigations of the effects of channel pathologies, channel blockers and ion concentrations on macroscopic activity, we formulate neural mass equations explicitly incorporating the single cell dynamics by using a bottom-up approach. The mean and variance of the firing rate and synaptic input distributions are modeled. The firing rate curve (F(I)-curve) is used as link between the single cell and macroscopic dynamics. We show that this model accurately reproduces the behavior of two populations of synaptically connected Hodgkin-Huxley neurons, also in non-steady state.     

An improved genetic algorithm based fuzzy-tuned neural network

This paper presents a fuzzy-tuned neural network, which is trained by an improved genetic algorithm (GA). The fuzzy-tuned neural network consists of a neural-fuzzy network and a modified neural network. In the modified neural network, a neuron model with two activation functions is used so that the degree of freedom of the network function can be increased. The neural-fuzzy network governs some of the parameters of the neuron model. It will be shown that the performance of the proposed fuzzy-tuned neural network is better than that of the traditional neural network with a similar number of parameters. An improved GA is proposed to train the parameters of the proposed network. Sets of improved genetic operations are presented. The performance of the improved GA will be shown to be better than that of the traditional GA. Some application examples are given to illustrate the merits of the proposed neural network and the improved GA.     

A path planning method based on the particle swarm optimization trained fuzzy neural network algorithm

Cluster Computing - The basic fuzzy neural network algorithm has slow convergence and large amount of calculation, so this paper designed a particle swarm optimization trained fuzzy neural network...     

A Lempel-Ziv complexity-based neural network pruning algorithm

This paper presents a pruning method for artificial neural networks (ANNs) based on the 'Lempel-Ziv complexity' (LZC) measure. We call this method the 'silent pruning algorithm' (SPA). The term 'silent' is used in the sense that SPA prunes ANNs without causing much disturbance during the network training. SPA prunes hidden units during the training process according to their ranks computed from LZC. LZC extracts the number of unique patterns in a time sequence obtained from the output of a hidden unit and a smaller value of LZC indicates higher redundancy of a hidden unit. SPA has a great resemblance to biological brains since it encourages higher complexity during the training process. SPA is similar to, yet different from, existing pruning algorithms. The algorithm has been tested on a number of challenging benchmark problems in machine learning, including cancer, diabetes, heart, card, iris, glass, thyroid, and hepatitis problems. We compared SPA with other pruning algorithms and we found that SPA is better than the 'random deletion algorithm' (RDA) which prunes hidden units randomly. Our experimental results show that SPA can simplify ANNs with good generalization ability.     

A visual metaphor describing neural dynamics in schizophrenia

Background

In many scientific disciplines the use of a metaphor as an heuristic aid is not uncommon. A well known example in somatic medicine is the ‘defense army metaphor’ used to characterize the immune system. In fact, probably a large part of the everyday work of doctors consists of ‘translating’ scientific and clinical information (i.e. causes of disease, percentage of succes versus risk of side-effects) into information tailored to the needs and capacities of the individual patient. The ability to do so in an effective way is at least partly what makes a clinician a good communicator. Schizophrenia is a severe psychiatric disorder which affects approximately 1% of the population. Over the last two decades a large amount of molecular-biological, imaging and genetic data have been accumulated regarding the biological underpinnings of schizophrenia. However, it remains difficult to understand how the characteristic symptoms of schizophrenia such as hallucinations and delusions are related to disturbances on the molecular-biological level. In general, psychiatry seems to lack a conceptual framework with sufficient explanatory power to link the mental- and molecular-biological domains.

Methodology/Principal Findings

Here, we present an essay-like study in which we propose to use visualized concepts stemming from the theory on dynamical complex systems as a ‘visual metaphor’ to bridge the mental- and molecular-biological domains in schizophrenia. We first describe a computer model of neural information processing; we show how the information processing in this model can be visualized, using concepts from the theory on complex systems. We then describe two computer models which have been used to investigate the primary theory on schizophrenia, the neurodevelopmental model, and show how disturbed information processing in these two computer models can be presented in terms of the visual metaphor previously described. Finally, we describe the effects of dopamine neuromodulation, of which disturbances have been frequently described in schizophrenia, in terms of the same visualized metaphor.

Conclusions/Significance

The conceptual framework and metaphor described offers a heuristic tool to understand the relationship between the mental- and molecular-biological domains in an intuitive way. The concepts we present may serve to facilitate communication between researchers, clinicians and patients.     

A genome alignment algorithm based on compression

Background  

Traditional genome alignment methods consider sequence alignment as a variation of the string edit distance problem, and perform alignment by matching characters of the two sequences. They are often computationally expensive and unable to deal with low information regions. Furthermore, they lack a well-principled objective function to measure the performance of sets of parameters. Since genomic sequences carry genetic information, this article proposes that the information content of each nucleotide in a position should be considered in sequence alignment. An information-theoretic approach for pairwise genome local alignment, namely XMAligner, is presented. Instead of comparing sequences at the character level, XMAligner considers a pair of nucleotides from two sequences to be related if their mutual information in context is significant. The information content of nucleotides in sequences is measured by a lossless compression technique.     

基于两级分段式算法的卷积神经网络的浮游有孔虫自动鉴定*

有孔虫个体微小、数量众多、地理分布广、演化迅速, 是记录海洋沉积环境的重要载体, 在海相生物地层划分和对比中具有十分重要的作用。因有孔虫属种众多, 传统的属种鉴定需要经验丰富的专业人员进行人工鉴定且耗时较长, 此外人工鉴定古生物面临人才匮乏和工作量大等问题。卷积神经网络在计算机视觉领域的应用可较好的解决上述问题。利用古生物专家对中新世浮游有孔虫化石标注为指导, 根据有孔虫化石不同方向的视角分类, 结合卷积神经网络算法, 开发了有孔虫化石图像识别系统。研究发现, 通过有孔虫化石腹视、缘视和背视角度分类, 采取两级分段式鉴定算法对中新世浮游有孔虫属一级进行识别, 属一级鉴定准确率达到82%左右。     

基于人工神经网络-遗传算法的樟芝发酵培养基优化

采用优化模型对药用丝状真菌樟芝的复杂发酵过程进行建模,并获得最优发酵培养基组成.对樟芝发酵过程中的形态变化过程进行了观察,并分别采用人工神经网络(ANN)和响应面法(RSM)对樟芝发酵过程进行建模,同时采用遗传算法(GA)优化了发酵培养基组成.结果表明,ANN模型比RSM模型具有更好的实验数据拟合能力和预测能力,GA计算得到樟芝生物量理论最大值为6.2 g/L,并获得发酵最佳接种量及培养基组成:孢子浓度1.76× 105个/mL,葡萄糖29.1 g/L,蛋白胨9.4 g/L,黄豆粉2.8 g/L.在最佳培养条件下,樟芝生物量为(6.1±0.2)g/L.基于ANN-GA的优化方法可用于优化其他丝状真菌的复杂发酵过程,从而获得生物量或活性代谢产物.     

A modified error backpropagation algorithm for complex-value neural networks

The complex-valued backpropagation algorithm has been widely used in fields of dealing with telecommunications, speech recognition and image processing with Fourier transformation. However, the local minima problem usually occurs in the process of learning. To solve this problem and to speed up the learning process, we propose a modified error function by adding a term to the conventional error function, which is corresponding to the hidden layer error. The simulation results show that the proposed algorithm is capable of preventing the learning from sticking into the local minima and of speeding up the learning.     

A novel neural response algorithm for protein function prediction

Background

Large amounts of data are being generated by high-throughput genome sequencing methods. But the rate of the experimental functional characterization falls far behind. To fill the gap between the number of sequences and their annotations, fast and accurate automated annotation methods are required. Many methods, such as GOblet, GOFigure, and Gotcha, are designed based on the BLAST search. Unfortunately, the sequence coverage of these methods is low as they cannot detect the remote homologues. Adding to this, the lack of annotation specificity advocates the need to improve automated protein function prediction.

Results

We designed a novel automated protein functional assignment method based on the neural response algorithm, which simulates the neuronal behavior of the visual cortex in the human brain. Firstly, we predict the most similar target protein for a given query protein and thereby assign its GO term to the query sequence. When assessed on test set, our method ranked the actual leaf GO term among the top 5 probable GO terms with accuracy of 86.93%.

Conclusions

The proposed algorithm is the first instance of neural response algorithm being used in the biological domain. The use of HMM profiles along with the secondary structure information to define the neural response gives our method an edge over other available methods on annotation accuracy. Results of the 5-fold cross validation and the comparison with PFP and FFPred servers indicate the prominent performance by our method. The program, the dataset, and help files are available at http://www.jjwanglab.org/NRProF/.

     

Wind power prediction based on neural network with optimization of adaptive multi-group salp swarm algorithm

Salp swarm algorithm (SSA) is a swarm intelligence algorithm inspired by the swarm behavior of salps in oceans. In this paper, a adaptive multi-group salp swarm algorithm (AMSSA) with three new communication strategies is presented. Adaptive multi-group mechanism is to evenly divide the initial population into several subgroups, and then exchange information among subgroups after each adaptive iteration. Communication strategy is also an important part of adaptive multi-group mechanism. This paper proposes three new communication strategies and focuses on promoting the performance of SSA. These measures significantly improve the cooperative ability of SSA, accelerate convergence speed, and avoid easily falling into local optimum. And the benchmark functions confirm that AMSSA is better than the original SSA in exploration and exploitation. In addition, AMSSA is combined with prediction of wind power based on back propagation (AMSSA-BP) neural network. The simulation results show that the AMSSA-BP neural network prediction model can achieve a better prediction effect of wind power.

     

Singular value decomposition learning on double Stiefel manifold

The aim of this paper is to present a unifying view of four SVD-neural-computation techniques found in the scientific literature and to present some theoretical results on their behavior. The considered SVD neural algorithms are shown to arise as Riemannian-gradient flows on double Stiefel manifold and their geometric and dynamical properties are investigated with the help of differential geometry.     

Few-sweep estimation of evoked potential based on a generalized subspace approach

A generalized subspace approach is proposed for single channel brain evoked potential (EP) extraction from background electroencephalogram (EEG) signal. The method realizes the optimum estimate of EP signal from the observable noisy signal. The underlying principle is to project the signal and noise into signal and noise coefficient subspace respectively by applying projection matrix at first. Secondly, coefficient weighting matrix is achieved based on the autocorrelation matrices of the noise and the noisy signal. With the coefficient weighting matrix, we can remove the noise projection coefficients and estimate the signal ones. EP signal is then obtained by averaging the signals estimated with the reconstruction matrix. Given different signal-to-noise ratio (SNR) conditions, the algorithm can estimate the EP signal with only two sweeps observable noisy signals. Our approach is shown to have excellent capability of estimating EP signal even in poor SNR conditions. The interference of spontaneous EEG has been eliminated with significantly improved SNR. The simulation results have demonstrated the effectiveness and superior performance of the proposed method.     

A real time learning algorithm for recurrent analog neural networks

A new learning algorithm is described for a general class of recurrent analog neural networks which ultimately settle down to a steady state. Recently, Pineda (Pineda 1987; Almeida 1987; Ikeda et al. 1988) has introduced a learning rule for the recurrent net in which the connection weights are adjusted so that the distance between the stable outputs of the current system and the desired outputs will be maximally decreased. In this method, many cycles are needed in order to get a target system. In each cycle, the recurrent net is run until it reaches a stable state. After that, the weight change is calculated by using a linearized recurrent net which receives the current error of the system as a bias input. In the new algorithm the weights are changed so that the total error of neuron outputs over the entire trajectory is minimized. The weights are adjusted in real time when the network is running. In this method, the trajectory to the target system can be controlled, whereas Pineda's algorithm only controls the position of the fixed point. The relation to the back propagation method (Hinton et al. 1986) is also discussed.     

A neural network algorithm for the multiple traveling salesmen problem

We developed an efficient neural network algorithm for solving the Multiple Traveling Salesmen Problem (MTSP). A new transformation of the N-city M-salesmen MTSP to the standard Traveling Salesmen Problem (TSP) is introduced. The transformed problem is represented by an expanded version of Hopfield-Tank's neuromorphic city-position map with (N + M-1)-cities and a single fictitious salesmen. The dynamic model associated with the problem is based on the Basic Differential Multiplier Method (BDMM) [26] which evaluates Lagrange multipliers simultaneously with the problem's state variables. The algorithm was successfully tested on many problems with up to 30 cities and five salesmen. In all test cases, the algorithm always converged to valid solutions. The great advantage of this kind of algorithm is that it can provide solutions to complex decision making problems directly by solving a system of ordinary differential equations. No learning steps, logical if statements or adjusting of parameters are required during the computation. The algorithm can therefore be implemented in hardware to solve complex constraint satisfaction problems such as the MTSP at the speed of analog silicon VLSI devices or possibly future optical neural computers.     

Scaling the neural TSP algorithm

Recent attempts to find a procedure for scaling neural shortest path computations have failed, leading to speculation that the basic optimization method is unreliable and without justification. We report the first effective procedures for scaling such algorithms and demonstrate their validity. Independently of the scaling, an unc onventional approach to neural simulation is described which surprisingly has much the same effect as simulated annealing, but without the need for adjustable run-time parameters (such as temperature).     

TBC: A clustering algorithm based on prokaryotic taxonomy

High-throughput DNA sequencing technologies have revolutionized the study of microbial ecology. Massive sequencing of PCR amplicons of the 16S rRNA gene has been widely used to understand the microbial community structure of a variety of environmental samples. The resulting sequencing reads are clustered into operational taxonomic units that are then used to calculate various statistical indices that represent the degree of species diversity in a given sample. Several algorithms have been developed to perform this task, but they tend to produce different outcomes. Herein, we propose a novel sequence clustering algorithm, namely Taxonomy-Based Clustering (TBC). This algorithm incorporates the basic concept of prokaryotic taxonomy in which only comparisons to the type strain are made and used to form species while omitting full-scale multiple sequence alignment. The clustering quality of the proposed method was compared with those of MOTHUR, BLASTClust, ESPRIT-Tree, CD-HIT, and UCLUST. A comprehensive comparison using three different experimental datasets produced by pyrosequencing demonstrated that the clustering obtained using TBC is comparable to those obtained using MOTHUR and ESPRIT-Tree and is computationally efficient. The program was written in JAVA and is available from .