Search

Abstract Significant interest exists in establishing synergistic research in bioinformatics, systems biology and intelligent computing. Supported by the United States National Science Foundation (NSF), International Society of Intelligent Biological Medicine (http://www.ISIBM.org), International Journal of Computational Biology and Drug Design (IJCBDD) and International Journal of Functional Informatics and Personalized Medicine, the ISIBM International Joint Conferences on Bioinformatics, Systems Biology and Intelligent Computing (ISIBM IJCBS 2009) attracted more than 300 papers and 400 researchers and medical doctors world-wide. It was the only inter/multidisciplinary conference aimed to promote synergistic research and education in bioinformatics, systems biology and intelligent computing. The conference committee was very grateful for the valuable advice and suggestions from honorary chairs, steering committee members and scientific leaders including Dr. Michael S. Waterman (USC, Member of United States National Academy of Sciences), Dr. Chih-Ming Ho (UCLA, Member of United States National Academy of Engineering and Academician of Academia Sinica), Dr. Wing H. Wong (Stanford, Member of United States National Academy of Sciences), Dr. Ruzena Bajcsy (UC Berkeley, Member of United States National Academy of Engineering and Member of United States Institute of Medicine of the National Academies), Dr. Mary Qu Yang (United States National Institutes of Health and Oak Ridge, DOE), Dr. Andrzej Niemierko (Harvard), Dr. A. Keith Dunker (Indiana), Dr. Brian D. Athey (Michigan), Dr. Weida Tong (FDA, United States Department of Health and Human Services), Dr. Cathy H. Wu (Georgetown), Dr. Dong Xu (Missouri), Drs. Arif Ghafoor and Okan K Ersoy (Purdue), Dr. Mark Borodovsky (Georgia Tech, President of ISIBM), Dr. Hamid R. Arabnia (UGA, Vice-President of ISIBM), and other scientific leaders. The committee presented the 2009 ISIBM Outstanding Achievement Awards to Dr. Joydeep Ghosh (UT Austin), Dr. Aidong Zhang (Buffalo) and Dr. Zhi-Hua Zhou (Nanjing) for their significant contributions to the field of intelligent biological medicine.

Effective partitioning of the feature space for high classification accuracy with due attention to rare class members is often a difficult task. In this paper, the border vector detection and adaptation (BVDA) algorithm is proposed for this purpose. The BVDA consists of two parts. In the first part of the algorithm, some specially selected training samples are assigned as initial reference vectors called border vectors. In the second part of the algorithm, the border vectors are adapted by moving them toward the decision boundaries. At the end of the adaptation process, the border vectors are finalized. The method next uses the minimum distance to border vector rule for classification. In supervised learning, the training process should be unbiased to reach more accurate results in testing. In the BVDA, decision region borders are related to the initialization of the border vectors and the input ordering of the training samples. Consensus strategy can be applied with cross validation to reduce these dependencies. The performance of the BVDA and consensual BVDA were studied in comparison to other classification algorithms including neural network with backpropagation learning, support vector machines, and some statistical classification techniques. Index Terms--Border vector detection and adaptation (BVDA), consensual classification, data classification, decision region borders, remote sensing.

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2006.03.040 Byline: Hoi-Ming Chi (a), Okan K. Ersoy (a), Herbert Moskowitz (b), Jim Ward (b) Keywords: Supply chain management; Support vector machines; Genetic algorithms; Machine learning; Decision support systems Abstract: Using a supply chain network, we demonstrate the feasibility, viability, and robustness of applying machine learning and genetic algorithms to respectively model, understand, and optimize such data intensive environments. Deployment of these algorithms, which learn from and optimize data, can obviate the need to perform more complex, expensive, and time consuming design of experiments (DOE), which usually disrupt system operations. We apply and compare the behavior and performance of the proposed machine learning algorithms to that obtained via DOE in a simulated Vendor Managed Replenishment system, developed for an actual firm. The results show that the models resulting from the proposed algorithms had strong explanatory and predictive power, comparable to that of DOE. The optimal system settings and profit were also similar to that obtained from DOE. The virtues of using machine learning and evolutionary algorithms to model and optimize data rich environments thus seem promising because they are automatic, involving little human intervention and expertise. We believe and are exploring how they can be made adaptive to improve parameter estimates with increasing data, as well as seamlessly detecting system (and therefore model) changes, thus being capable of recursively updating and reoptimizing a modified or new model. Author Affiliation: (a) School of Electrical and Computer Engineering, Purdue University, Electrical Engineering Building, 465 Northwestern Avenue, West Lafayette, IN 47907-2035, USA (b) The Krannert School of Management, Purdue University, 403 West State Street, West Lafayette, IN 47907-2056, USA Article History: Received 30 October 2004; Accepted 1 March 2006

Information technology (IT) has been the driver of increased productivity in the manufacturing and service sectors, bringing real-time information to decision makers and process owners to improve process behavior and […]

A new learning system called a statistical self-organizing learning system (SSOLS), combining functional-link neural networks, statistical hypothesis testing, and self-organization of a number of enhancement nodes, is introduced for remote sensing applications. Its structure consists of two stages, a mapping stage and a learning stage. The input training vectors are initially mapped to the enhancement vectors in the mapping stage by multiplying with a random matrix, followed by pointwise nonlinear transformations. Starting with only one enhancement node, the enhancement layer incrementally adds an extra node in each iteration. The optimum dimension of the enhancement layer is determined by using an efficient leave-one-out cross-validation method. In this way, the number of enhancement nodes is also learned automatically. A t-test algorithm can also be applied to the mapping stage to mitigate the effect of overfitting and to further reduce the number of enhancement nodes required, resulting in a more compact network. In the learning stage, both the input vectors and the enhancement vectors are fed into a least squares learning module to obtain the estimated output vectors. This is made possible by choosing the output layer linear. In addition, several SSOLSs can be trained independently in parallel to form a consensual SSOLS, whose final output is a linear combination of the outputs of each SSOLS module. The SSOLS is simple, fast to compute, and suitable for remote sensing applications, especially with hyperspectral image data of high dimensionality. Index Terms--Artificial neural networks, hyperspectral image classification, machine learning, overfitting, support vector machine (SMV), t-test.

In multi-objective optimization problems, the objective space of fitness functions has a close relationship with the solution space. Extracting the optimal direction and optimal parameter information are very useful for the optimization process. This paper proposes multi-objective differential evolution algorithm with a clustering based objective space division and parameter adaptation (MODECD). L metric matrix based optimal strategy is used to split the objective space into sub-spaces and to extract the optimal directions. A fitness value based parameter adaptation and mutation strategy are used to extract the optimal strategy information. The results with 20 benchmark tests show the competitiveness of the MODECD algorithm in both convergence speed and diversity of solution approximating the Pareto front. In addition, MODECD is used to optimize the fermentation process of sodium gluconate as an example of its superior performance in solving real-world problems. [ABSTRACT FROM AUTHOR]

Classification accuracy improvement of neural network classifiers using unlabeled testing data is presented. In order to increase the classification accuracy without increasing the number of training data, the network makes use of testing data along with training data for learning. It is shown that including the unlabeled samples from underrepresented classes in the training set improves the classification accuracy of some of the classes during supervised-unsupervised learning.

A new type of a neural-network architecture, the parallel consensual neural network (PCNN), is introduced and applied in classification/data fusion of multisource remote sensing and geographic data. The PCNN architecture is based on statistical consensus theory and involves using stage neural networks with transformed input data. The input data are transformed several times and the different transformed data are used as if they were independent inputs. The independent inputs are first classified using the stage neural networks. The output responses from the stage networks are then weighted and combined to make a consensual decision. In this paper, optimization methods are used in order to weight the outputs from the stage networks. Two approaches are proposed to compute the data transforms for the PCNN, one for binary data and another for analog data. The analog approach uses wavelet packets. The experimental results obtained with the proposed approach show that the PCNN outperforms both a conjugate-gradient backpropagation neural network and conventional statistical methods in terms of overall classification accuracy of test data. Index Terms - Consensus theory, wavelet packets, accuracy, classification, probability density estimation, statistical pattern recognition, time-frequency analysis, data fusion.

Parallel, self-organizing, hierarchical neural networks (PSHNN's) are multistage networks in which stages operate in parallel rather than in series during testing. Each stage can be any particular type of network. Previous PSHNN's assume quantized, say, binary outputs. A new type of PSHNN is discussed such that the outputs are allowed to be continuous-valued. The performance of the resulting networks is tested in the problem of predicting speech signal samples from past samples. Three types of networks in which the stages are learned by the delta rule, sequential least-squares, and the backpropagation (BP) algorithm, respectively, are described. In all cases studied, the new networks achieve better performance than linear prediction. A revised BP algorithm is discussed for learning input nonlinearities. When the BP algorithm is to be used, better performance is achieved when a single BP network is replaced by a PSHNN of equal complexity in which each stage is a BP network of smaller complexity than the single BP network.

A scheme for the design of diffractive phase elements (DPE's) that integrates several optical functions is presented in a consistent sense based on the general theory of amplitude-phase retrieval and the Yang-Gu algorithm [Appl. Opt. 33, 209 (1994)!. We extend the original Yang-Gu algorithm to treat a system illuminated by a beam of incident light whose components are at different wavelengths, and a set of equations for determining the phase distribution of the DPE is derived. The profile of a surface-relief DPE can be designed with an iterative algorithm. Numerical simulations are carried out for the design of one-dimensional DPE's capable of both demultiplexing different wavelength components and focusing each partial wave at predetermined positions. The influence of the extension of sampling points in the DPE's from ideal geometric points to physical spots on design results is also investigated. The numerical simulation results show that the new algorithm can be used successfully to design the desired DPE's. It is therefore expected to be useful in the design of DPE's for micro-optical systems.

A multistage parallel algorithm with iterative processing is discussed for the processing of information in diffraction tomography. The algorithm is based on matrix partitioning, which results in mostly parallel stages of processing. Each successive stage is designed to minimize the remaining error. The process is iterated until convergence. The major advantages of the multistage algorithm are the reduced computational time from faster convergence as compared with a single-stage iterative algorithm, further reduction of computation time if the stages are implemented mostly in parallel, and better performance in terms of reduced reconstruction error.

A detailed comparison of the original Gerchberg-Saxton and the Yang-Gu algorithms for the reconstruction of model images from two intensity measurements in a nonunitary transform system is presented. The Yang-Gu algorithm is a generalization of the Gerchberg-Saxton algorithm and is effective in solving the general amplitude-phase-retrieval problem in any linear unitary or nonunitary transform system. For a unitary transform system the Yang-Gu algorithm is identical to the Gerchberg-Saxton algorithm. The reconstruction of images from data corrupted with random noise is also investigated. The simulation results show that the Yang-Gu algorithm is relatively insensitive to the presence of noise in data. In all cases studied the Yang-Gu algorithm always resulted in a highly accurate recovered phase.

A new neural network learning algorithm with competitive learning and multiple safe rejection schemes is proposed in the context of parallel, self-organizing, hierarchical neural networks (PSHNN). After reference vectors are computed using competitive learning in a stage of PSHNN, the safe rejection schemes are constructed for reference vectors. The purpose of safe rejection schemes is to reject the input vectors which are hard to classify. The next stage neural network is trained with the nonlinearly transformed values of only those training vectors that were rejected in the previous stage neural network. Two different kinds of safe rejection schemes, RADPN and RAD, are developed and used together. Experimental results comparing the performance of the proposed algorithms with those of backpropagation and the PSHNN with the delta rule learning algorithm are discussed. The proposed learning network produced higher classification accuracy and much faster learning. The classification accuracies of two methods for learning the reference vectors were compared. When the reference vectors are computed separately for each class (Method II), higher classification accuracy was obtained as compared to the method in which the reference vectors are computed together for all the classes (Method I). This conclusion has to do with rejection of hard vectors, and is the opposite of what is normally expected. In addition, Method II has the advantage of parallelism by which the reference vectors for all the classes can be computed simultaneously.

Parallel, self-organizing, hierarchical neural networks (PSHNN's) involve a number of stages with error detection at the end of each stage, rejection of error-causing vectors, which are then fed into the next stage after a nonlinear transformation. The stages operate in parallel during testing. Statistical properties and the mechanisms of vector rejection of the PSHNN are discussed in comparison to the maximum likelihood method and the backpropagation network. The PSHNN is highly fault tolerant and robust against errors in the weight values due to the adjustment of the error detection bounds to compensate errors in the weight values. These properties are exploited to develop architectures for programmable implementations in which the programmable parts are reduced to on-off or bipolar switching operations for bulk computations and attenuators for pointwise operations.

A general framework containing four categories, similar to that of the generalized discrete Fourier transform, is obtained for the discrete Hartley transform (DHT). The derived algorithms are suitable for digital filter bank design, signal modeling, convolution calculation and speedy computation of the DHT in cases of complex data points. A discussion on the use of the generalized transforms for skew-convolution computation is also given.

An analysis of discrete Fourier preprocessing transforms (DFPT) shows that these, along with discrete and real discrete Fourier transforms, can be computed for in the second stage of two-stage Fourier algorithms. DFPTs can also be derived by using skew-circular correlations (SCC) followed by fast SCC algorithms although these may introduce complex elements into the DFPTs' simpler matrices. However, using T(sub 25) to implement other DFPTs avoids this problem since T(sub 25) is the simplest DFPT to implement and SCCs have few nonzero elements.