Zhidong Deng

Collective Behavior of a Small-World Recurrent Neural System With Scale-Free Distribution

This paper proposes a scale-free highly clustered echo state network (SHESN). We designed the SHESN to include a naturally evolving state reservoir according to incremental growth rules that account for the following features: (1) short characteristic path length, (2) high clustering coefficient, (3) scale-free distribution, and (4) hierarchical and distributed architecture. This new state reservoir contains a large number of internal neurons that are sparsely interconnected in the form of domains. Each domain comprises one backbone neuron and a number of local neurons around this backbone. Such a natural and efficient recurrent neural system essentially interpolates between the completely regular Elman network and the completely random echo state network (ESN) proposed by Jaeger We investigated the collective characteristics of the proposed complex network model. We also successfully applied it to challenging problems such as the Mackey-Glass (MG) dynamic system and the laser time-series prediction. Compared to the ESN, our experimental results show that the SHESN model has a significantly enhanced echo state property and better performance in approximating highly complex nonlinear dynamics. In a word, this large scale dynamic complex network reflects some natural characteristics of biological neural systems in many aspects such as power law, small-world property, and hierarchical architecture. It should have strong computing power, fast signal propagation speed, and coherent synchronization.

Computational prediction of novel non-coding RNAs in Arabidopsis thaliana.

BackgroundNon-coding RNA (ncRNA) genes do not encode proteins but produce functional RNA molecules that play crucial roles in many key biological processes. Recent genome-wide transcriptional profiling studies using tiling arrays in organisms such as human and Arabidopsis have revealed a great number of transcripts, a large portion of which have little or no capability to encode proteins. This unexpected finding suggests that the currently known repertoire of ncRNAs may only represent a small fraction of ncRNAs of the organisms. Thus, efficient and effective prediction of ncRNAs has become an important task in bioinformatics in recent years. Among the available computational methods, the comparative genomic approach seems to be the most powerful to detect ncRNAs. The recent completion of the sequencing of several major plant genomes has made the approach possible for plants.ResultsWe have developed a pipeline to predict novel ncRNAs in the Arabidopsis (Arabidopsis thaliana) genome. It starts by comparing the expressed intergenic regions of Arabidopsis as provided in two whole-genome high-density oligo-probe arrays from the literature with the intergenic nucleotide sequences of all completely sequenced plant genomes including rice (Oryza sativa), poplar (Populus trichocarpa), grape (Vitis vinifera), and papaya (Carica papaya). By using multiple sequence alignment, a popular ncRNA prediction program (RNAz), wet-bench experimental validation, protein-coding potential analysis, and stringent screening against various ncRNA databases, the pipeline resulted in 16 families of novel ncRNAs (with a total of 21 ncRNAs).ConclusionIn this paper, we undertake a genome-wide search for novel ncRNAs in the genome of Arabidopsis by a comparative genomics approach. The identified novel ncRNAs are evolutionarily conserved between Arabidopsis and other recently sequenced plants, and may conduct interesting novel biological functions.

Anchor-Free Localization Method for Mobile Targets in Coal Mine Wireless Sensor Networks

Severe natural conditions and complex terrain make it difficult to apply precise localization in underground mines. In this paper, an anchor-free localization method for mobile targets is proposed based on non-metric multi-dimensional scaling (Multi-dimensional Scaling: MDS) and rank sequence. Firstly, a coal mine wireless sensor network is constructed in underground mines based on the ZigBee technology. Then a non-metric MDS algorithm is imported to estimate the reference nodes’ location. Finally, an improved sequence-based localization algorithm is presented to complete precise localization for mobile targets. The proposed method is tested through simulations with 100 nodes, outdoor experiments with 15 ZigBee physical nodes, and the experiments in the mine gas explosion laboratory with 12 ZigBee nodes. Experimental results show that our method has better localization accuracy and is more robust in underground mines.

Trading strategy design in financial investment through a turning points prediction scheme

Turning points prediction has long been a tough task in the field of time series analysis due to its strong nonlinearity, and thus has attracted many research efforts. In this study, the turning points prediction (TPP) framework is presented and further employed to develop a novel trading strategy designing approach to financial investment. The TPP framework is a machine learning-based solution incorporating chaotic dynamic analysis and neural network modeling. It works on the ground of a nonlinear mapping deduced in financial time series through chaotic analysis. An event characterization method is created in TTP framework to characterize trend patterns in ongoing financial time series. The main contributions of this paper are (1) it presents an ensemble learning based TPP framework, within which the nonlinear mapping is approximated by the ensemble artificial neural network (EANN) model with a new parameters learning algorithm; (2) a genetic algorithm (GA) based threshold optimization procedure is described with a newly defined performance measure, named TpMSE, which is used as a cost function; and (3) a trading strategy designing approach is proposed based on the TPP framework. The proposed approach was applied to the two real-world financial time series, i.e., an individual stock quote time series and the Dow Jones Industrial Average (DJIA) index time series. Experimental results show that the proposed approach can help investors make profitable decisions.

Analysis of power spectrum and 1/f type power law in a complex computer network model

Abstract The computer network, as a complex nonlinear system, exhibits plenty of nonlinear complexities phenomena. More and more researchers have been attracted to these exciting dynamic behaviors. In general, computer simulation studies are widely used in the researches on network dynamics. Unfortunately, too simple and regular topologies are adopted in these network models. It inevitably results in the fact that some important complexities behaviors in a real computer network may not emerge in most of computer experimental studies. In this paper, the phase transition phenomena between the packet creation rate and the lifetime of packet, on basis of more realistic computer network models composed of the Transit–Stub topology and the proposed mechanism of packet traffic, are explored. To illustrate the fluctuation of packets more clearly, a power spectrum is made and the 1/f type power law is found. Finally, simulation results and some research emphases in the future are discussed.

Cellular automaton modeling of computer network

The dramatically increasing growth of computer network has brought us more and more complex systems that reveals plenty of nonlinear complex phenomena. On the other hand, cellular automaton model used to be in the study of a variety of nonlinear and spatially extended systems. This paper proposes a simplified cellular model for computer network, namely the NaSch network model, which is originated at the NaSch model of road traffic. Typically, the NaSch network model is a one-dimension cellular automaton and consists of two kinds of cells, i.e. node cell and link cell. In this paper, the node cell stands for switch nodes in computer network, such as routers and switchers, while the link cell is the abstract of the lines of communication among these switch nodes. The simulation results show that this model indeed captures some of properties of flux of computer networks. The space–time plots illustrate that the randomization in this modeling plays an important role in the emerging of self-organization of congestion. It also demonstrates that the density of packet is another key factor to having influence on congestion. On the contrary, the boundary condition has few contribution to our model.  2002 Elsevier Science B.V. All rights reserved.

An Accurate GPS-IMU/DR Data Fusion Method for Driverless Car Based on a Set of Predictive Models and Grid Constraints

A high-performance differential global positioning system (GPS) receiver with real time kinematics provides absolute localization for driverless cars. However, it is not only susceptible to multipath effect but also unable to effectively fulfill precise error correction in a wide range of driving areas. This paper proposes an accurate GPS–inertial measurement unit (IMU)/dead reckoning (DR) data fusion method based on a set of predictive models and occupancy grid constraints. First, we employ a set of autoregressive and moving average (ARMA) equations that have different structural parameters to build maximum likelihood models of raw navigation. Second, both grid constraints and spatial consensus checks on all predictive results and current measurements are required to have removal of outliers. Navigation data that satisfy stationary stochastic process are further fused to achieve accurate localization results. Third, the standard deviation of multimodal data fusion can be pre-specified by grid size. Finally, we perform a lot of field tests on a diversity of real urban scenarios. The experimental results demonstrate that the method can significantly smooth small jumps in bias and considerably reduce accumulated position errors due to DR. With low computational complexity, the position accuracy of our method surpasses existing state-of-the-arts on the same dataset and the new data fusion method is practically applied in our driverless car.

Behaviours of networks with different topologies and protocols

In researches on computer network, simulation experiments for various different topological structures are used widely. But it is still a confusing problem that how these topologies have influence upon the results in these researches. In this paper, several popular topological structures, such as the Transit-Stub, the two-dimension lattice, and the BRITE, are adopted in simulation experiments so as to study their contributions to the results of simulations. It is discovered that network models based on different topological structures exhibit similar behaviours as a whole. In addition, dynamic rules are considered as one of the main sources of the complex phenomena occurred in computer networks. Because a network protocol covers the most aspect of dynamic rules in computer networks, it plays an important role in a dynamic evolutive process. This paper uses the connectionless UDP protocol and the connection-oriented TCP one to investigate such contributions. As a result, different behaviours are observed in the simulation. Finally, a conclusion remark is drawn.

Accelerating Convolutional Neural Networks with Dominant Convolutional Kernel and Knowledge Pre-regression

Aiming at accelerating the test time of deep convolutional neural networks (CNNs), we propose a model compression method that contains a novel dominant kernel (DK) and a new training method called knowledge pre-regression (KP). In the combined model DK\(^2\)PNet, DK is presented to significantly accomplish a low-rank decomposition of convolutional kernels, while KP is employed to transfer knowledge of intermediate hidden layers from a larger teacher network to its compressed student network on the basis of a cross entropy loss function instead of previous Euclidean distance. Compared to the latest results, the experimental results achieved on CIFAR-10, CIFAR-100, MNIST, and SVHN benchmarks show that our DK\(^2\)PNet method has the best performance in the light of being close to the state of the art accuracy and requiring dramatically fewer number of model parameters.

Complex Systems Modeling Using Scale-Free Highly-Clustered Echo State Network

Inspired by the universal laws governing different kinds of complex networks, we propose a scale-free highly-clustered echo state network (SHESN). Different from echo state network (ESN), the state reservoir of the SHESN is generated by natural growth rules and eventually forms a complex network with small-world, scale-free properties, and hierarchically distributed structure. We implemented a large-scale SHESN with 3,000 internal neurons and applied it to modeling the pH-neutralization process. Simulation results showed the superior performance of SHESN. Furthermore, we analyzed the natural characteristics of the SHESN and discussed our growth rules and the new state reservoir from a brain functional network perspective.