Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop

24-26 Sept. 1997

Filter Results

Displaying Results 1 - 25 of 74
  • Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop

    Publication Year: 1997
    Request permission for commercial reuse | PDF file iconPDF (311 KB)
    Freely Available from IEEE
  • A neural network equalizer with the fuzzy decision learning rule

    Publication Year: 1997, Page(s):551 - 559
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (377 KB)

    We propose a neural network equalizer with a fuzzy decision learning rule based on the generalized probabilistic descent algorithm with the minimum decision error formulation. The neural network used is the multi-layer perceptron. It is shown that the decision region overlapped by noise can be overcome by the use of a fuzzy decision learning rule based on the generalized probabilistic descent algo... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Author index

    Publication Year: 1997, Page(s):665 - 667
    Request permission for commercial reuse | PDF file iconPDF (85 KB)
    Freely Available from IEEE
  • Reducing false alarm risk in transient signal classification

    Publication Year: 1997, Page(s):112 - 120
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (576 KB)

    We address the problem of autonomous decision making in classification of radioastronomy transient signals on spectrograms from spacecraft. It is known that the assessment of the decision process can be divided into acceptance of the classification, instant rejection of the current signal classification, or rejection of the entire classifier model. We propose to combine prediction and classificati... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Blind source separation and deconvolution by dynamic component analysis

    Publication Year: 1997, Page(s):456 - 465
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (444 KB)

    We derive new unsupervised learning rules for blind separation of mixed and convolved sources. These rules are nonlinear in the signals and thus exploit high-order spatiotemporal statistics to achieve separation. The derivation is based on a global optimization formulation of the separation problem, yielding a stable algorithm. Different rules are obtained from frequency- and time-domain optimizat... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Combined learning and use for classification and regression models

    Publication Year: 1997, Page(s):102 - 111
    Cited by:  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (504 KB)

    We show that the decision function of a radial basis function (RBF) classifier is equivalent in form to the Bayes-optimal discriminant associated with a special kind of mixture-based statistical model. The relevant mixture model is a type of “mixture of experts” model for which class labels, like continuous-valued features, are assumed to have been generated randomly, conditional on th... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Combining discriminant-based classifiers using the minimum classification error discriminant

    Publication Year: 1997, Page(s):365 - 374
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (492 KB)

    Focusing on classification problems, this paper presents a new method for linearly combining discriminant-based classifiers to improve classification performance, in the sense of the minimum classification errors. In our approach, the problem of estimating linear weights in combination is reformulated as the problem of designing a linear discriminant function using the minimum classification error... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recurrent canonical piecewise linear network: theory and application

    Publication Year: 1997, Page(s):446 - 455
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (492 KB)

    A recurrent canonical piecewise linear (RCPL) network is defined by combining the canonical piecewise linear function with the autoregressive moving average (ARMA) model such that an augmented input space is partitioned into regions where an ARMA model is used in each. Properties of the RCPL network are discussed. Particularly, it is shown that the RCPL function is a contractive mapping and is sta... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On-line adaptive algorithms in non-stationary environments using a modified conjugate gradient approach

    Publication Year: 1997, Page(s):316 - 325
    Cited by:  Papers (7)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (524 KB)

    In this paper we propose novel computationally efficient schemas for a large class of online adaptive algorithms with variable self-adaptive learning rates. The learning rate is adjusted automatically providing relatively fast convergence at early stages of adaptation while ensuring small final misadjustment for cases of stationary environments. For nonstationary environments, the algorithms propo... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Extracting the relevant delays in time series modelling

    Publication Year: 1997, Page(s):92 - 101
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (584 KB)

    In this contribution, we suggest a convenient way to use generalisation error to extract the relevant delays from a time-varying process, i.e. the delays that lead to the best prediction performance. We design a generalisation-based algorithm that takes its inspiration from traditional variable selection, and more precisely stepwise forward selection. The method is compared to other forward select... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Feature extraction approach to blind source separation

    Publication Year: 1997, Page(s):398 - 405
    Cited by:  Papers (9)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (472 KB)

    Local independent component analysis is formulated as a task involving the extraction of local geometric structure in the joint distribution. Because the geometrical structure of statistical independence is not well captured by statistical descriptions such as moments and cumulants, we use feature detection tools from image analysis to locate the local independent component coordinate system. The ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Training recurrent networks

    Publication Year: 1997, Page(s):355 - 364
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (596 KB)

    Training recurrent networks is generally believed to be a difficult task. Excessive training times and lack of convergence to an acceptable solution are frequently reported. In this paper we seek to explain the reason for this from a numerical point of view and show how to avoid problems when training. In particular we investigate ill-conditioning, the need for and effect of regularization and ill... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multichannel blind separation and deconvolution of sources with arbitrary distributions

    Publication Year: 1997, Page(s):436 - 445
    Cited by:  Papers (31)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (480 KB)

    Blind deconvolution and separation of linearly mixed and convolved sources is an important and challenging task for numerous applications. While several algorithms have shown promise in these tasks, these techniques may fail to separate signal mixtures containing both sub- and super-Gaussian distributed sources. In this paper, we present a simple and efficient extension of a family of algorithms t... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Wave propagation as a neural coupling mechanism: hardware for self-organizing feature maps and the representation of temporal sequences

    Publication Year: 1997, Page(s):306 - 315
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (532 KB)

    Wave propagation within a “cortex” of neurons is introduced as a neural coupling mechanism. Using this effect for the control of the neural learning process, the network generates self-organizing feature maps. Additionally, wave propagation is used to influence the neural competition in representing the input of the network. By this means the network is able to represent temporal aspec... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Interpretation of recurrent neural networks

    Publication Year: 1997, Page(s):82 - 91
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (560 KB)

    This paper addresses techniques for interpretation and characterization of trained recurrent nets for time series problems. In particular, we focus on assessment of effective memory and suggest an operational definition of memory. Further we discuss the evaluation of learning curves. Various numerical experiments on time series prediction problems are used to illustrate the potential of the sugges... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A neural network approach to blind source separation

    Publication Year: 1997, Page(s):486 - 495
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (524 KB)

    The problem of adapting linear multi-input-multi-output systems for unsupervised separation of linear mixtures of sources arises in a number of signal processing applications. In this paper we present a new single layer neural network in which information transfer maximization is equivalent to minimizing a cost function involving the well-known constant modulus criterion originally used in blind e... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • One-unit contrast functions for independent component analysis: a statistical analysis

    Publication Year: 1997, Page(s):388 - 397
    Cited by:  Papers (36)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (544 KB)

    The author (1997) introduced a large family of one-unit contrast functions to be used in independent component analysis (ICA). In this paper, the family is analyzed mathematically in the case of a finite sample. Two aspects of the estimators obtained using such contrast functions are considered: asymptotic variance, and robustness against outliers. An expression for the contrast function that mini... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Separable non-linear least-squares minimization-possible improvements for neural net fitting

    Publication Year: 1997, Page(s):345 - 354
    Cited by:  Papers (24)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (532 KB)

    Neural network minimization problems are often ill-conditioned and in this contribution two ways to handle this will be discussed. It is shown that a better conditioned minimization problem can be obtained if the problem is separated with respect to the linear parameters. This will increase the convergence speed of the minimization. The Levenberg-Marquardt minimization method is often concluded to... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Blind signal deconvolution by spatio-temporal decorrelation and demixing

    Publication Year: 1997, Page(s):426 - 435
    Cited by:  Papers (7)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (468 KB)

    We present a simple efficient local unsupervised learning algorithm for online adaptive multichannel blind deconvolution and separation of i.i.d. sources. Under mild conditions, there exits a stable inverse system so that the source signals can be exactly recovered from their convolutive mixtures. Based on the existence of the inverse filter, we construct a two-stage neural network which consists ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Combination of adaptive signal processing and neural classification using an extended backpropagation algorithm

    Publication Year: 1997, Page(s):296 - 305
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (412 KB)

    Beside the use of purely neural systems, the combination of preprocessing units and neural classifiers has been used for a variety of signal segmentation and classification tasks. Whereas this approach reduces the input dimensionality as well as the complexity of the classification problem, its performance crucially depends on a proper preprocessing scheme, i.e., feature extraction. In this contri... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A deterministic annealing approach to discriminative hidden Markov model design

    Publication Year: 1997, Page(s):266 - 275
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (524 KB)

    We present the problem of designing a classifier system based on hidden Markov models (HMMs) from a labeled training set with the objective of minimizing the rate of misclassification. To design the globally optimal recognizer, all the HMMs must be jointly optimized to minimize the number of mis-classified training patterns. This is a difficult design problem which we attack using the technique of... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Mixture of discriminative learning experts of constant sensitivity for automated cytology screening

    Publication Year: 1997, Page(s):152 - 161
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (504 KB)

    One practical objective in an automated cytology screening task is to obtain as high as possible specificity (the percentage of normal slides being classified as normal) while attaining acceptable (predefined) constant sensitivity. In this paper, we propose a new learning algorithm which continuously improves the specificity while maintaining constant sensitivity for pattern classification problem... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Texture analysis and artificial neural network for detection of clustered microcalcifications on mammograms

    Publication Year: 1997, Page(s):199 - 206
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (392 KB)

    Clustered microcalcifications on X-ray mammograms are an important sign in the detection of breast cancer. This paper quantitatively describes the usefulness of texture analysis methods for the detection of clustered microcalcifications on digitized mammograms. Comparative studies of texture analysis methods are performed for the proposed texture analysis method, called the surrounding region depe... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On estimation of nonlinear black-box models: how to obtain a good initialization

    Publication Year: 1997, Page(s):72 - 81
    Cited by:  Papers (6)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (492 KB)

    An algorithm to define and initialize nonlinear recurrent neural net models using linear models is described. From a modeling point of view it is natural to try linear models first and then continue with nonlinear models. The suggested method gives such an algorithm and the nonlinear recurrent model is defined as an extension of the linear model. This gives less problems with local minima compared... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive regularization of neural classifiers

    Publication Year: 1997, Page(s):24 - 33
    Cited by:  Papers (13)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (496 KB)

    We present a regularization scheme which iteratively adapts the regularization parameters by minimizing the validation error. It is suggested to use the adaptive regularization scheme in conjunction with optimal brain damage pruning to optimize the architecture and to avoid overfitting. Furthermore, we propose an improved neural classification architecture eliminating an inherent redundancy in the... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.