In this paper, we present a novel hybrid architecture for continuous speech recognition systems. It consists of a continuous HMM system extended by an arbitrary neural network that is used as a preprocessor that takes several frames of the feature vector as input to produce more discriminative featurevectors with respect to the underlying HMM system. This hybrid system is an extension of a state-of-the-art continuous HMM system, andin fact, it is the first hybrid system that really is capable ofoutperforming thesestandard systems with respect to the recognition accuracy. Experimental results show an relative error reduction of about 10% that we achieved on a remarkably good recognition system based on continuous HMMsfor the Resource Management 1OOO-word continuous speech recognition task.

Applications of Gaussian mixture models occur frequently in the fields of statistics and artificial neural networks. One of the key issues arising from any mixture model application is how to estimate theoptimum number of mixture components. This paper extends the Reversible-Jump Markov Chain Monte Carlo (MCMC) algorithm to the case of multivariate spherical Gaussian mixtures using a hierarchical prior model. Using this method the number of mixture components is no longer fixed but becomes a parameter ofthe model which we shall estimate. The Reversible-Jump MCMC algorithm is capable of moving between parameter subspaces whichcorrespond to models with different numbers of mixture components. As a result a sample from the full joint distribution of all unknown model parameters is generated. The technique is then demonstrated on a simulated example and a well known vowel dataset. 1 Introduction Applications of Gaussian mixture models regularly appear in the neural networks literature. One of their most common roles in the field of neural networks, is in the placement of centres in a radial basis function network.

The proof given here shows that for any finite, discrete transfer function used by the units of an RCC network, there are finite-state automata (FSA) that the network cannot model, no matter how many units are used. The proof also applies to continuous transfer functions with a finite number of fixed-points, such as sigmoid and radial-basis functions.

In this paper, we discuss regularisation in online/sequential learning algorithms.In environments where data arrives sequentially, techniques such as cross-validation to achieve regularisation or model selection are not possible. Further, bootstrapping to determine aconfidence level is not practical. To surmount these problems, a minimum variance estimation approach that makes use of the extended Kalman algorithm for training multi-layer perceptrons isemployed. The novel contribution of this paper is to show the theoretical links between extended Kalman filtering, Sutton's variable learning rate algorithms and Mackay's Bayesian estimation framework.In doing so, we propose algorithms to overcome the need for heuristic choices of the initial conditions and noise covariance matrices in the Kalman approach.

The prior can be obtained by placing prior distributions on the weights in a neural 494 P. W Goldberg, C. K. L Williams and C. M. Bishop network, although we would argue that it is perhaps more natural to place priors directly overfunctions. One tractable way of doing this is to create a Gaussian process prior. This has the advantage that predictions can be made from the posterior using only matrix multiplication for fixed hyperparameters and a global noise level. In contrast, for neural networks (with fixed hyperparameters and a global noise level) it is necessary to use approximations or Markov chain Monte Carlo (MCMC) methods. Rasmussen(1996) has demonstrated that predictions obtained with Gaussian processes are as good as or better than other state-of-the art predictors. In much of the work on regression problems in the statistical and neural networks literatures, it is assumed that there is a global noise level, independent of the input vector x. The book by Bishop (1995) and the papers by Bishop (1994), MacKay (1995) and Bishop and Qazaz (1997) have examined the case of input-dependent noise for parametric models such as neural networks.

We study online generalized linear regression with multidimensional outputs, i.e., neural networks with multiple output nodes but no hidden nodes. We allow at the final layer transfer functions such as the softmax functionthat need to consider the linear activations to all the output neurons. We use distance functions of a certain kind in two completely independent roles in deriving and analyzing online learning algorithms for such tasks. We use one distance function to define a matching loss function for the (possibly multidimensional) transfer function, which allows usto generalize earlier results from one-dimensional to multidimensional outputs.We use another distance function as a tool for measuring progress made by the online updates. This shows how previously studied algorithmssuch as gradient descent and exponentiated gradient fit into a common framework. We evaluate the performance of the algorithms usingrelative loss bounds that compare the loss of the online algoritm to the best off-line predictor from the relevant model class, thus completely eliminating probabilistic assumptions about the data.

Active data clustering is a novel technique for clustering of proximity datawhich utilizes principles from sequential experiment design in order to interleave data generation and data analysis. The proposed activedata sampling strategy is based on the expected value of information, a concept rooting in statistical decision theory. This is considered to be an important step towards the analysis of largescale datasets, because it offers a way to overcome the inherent data sparseness of proximity data.

We explore methods for incorporating prior knowledge about a problem at hand in Support Vector learning machines. We show that both invariances undergroup transfonnations and prior knowledge about locality in images can be incorporated by constructing appropriate kernel functions.

Bayesian methods have been successfully applied to regression and classification problems in multi-layer perceptrons. We present a novel application of Bayesian techniques to Radial Basis Function networks by developing a Gaussian approximation to the posterior distribution which, for fixed basis function widths, is analytic in the parameters. The setting of regularization constants by crossvalidation iswasteful as only a single optimal parameter estimate is retained. We treat this issue by assigning prior distributions to these constants, which are then adapted in light of the data under a simple re-estimation formula. 1 Introduction Radial Basis Function networks are popular regression and classification tools[lO]. For fixed basis function centers, RBFs are linear in their parameters and can therefore betrained with simple one shot linear algebra techniques[lO]. The use of unsupervised techniques to fix the basis function centers is, however, not generally optimal since setting the basis function centers using density estimation on the input data alone takes no account of the target values associated with that data. Ideally, therefore, we should include the target values in the training procedure[7, 3, 9]. Unfortunately, allowingcenters to adapt to the training targets leads to the RBF being a nonlinear function of its parameters, and training becomes more problematic. Most methods that perform supervised training of RBF parameters minimize the ·Present address: SNN, University of Nijmegen, Geert Grooteplein 21, Nijmegen, The Netherlands.

Voice: (818) 306-6144 FAX: (818) 306-6912 Content Areas: Applications (Stochastic Optimization),Model Selection Algorithms Abstract This paper considers the problem of learning the ranking of a set of alternatives based upon incomplete information (e.g., a limited number of observations). We describe two algorithms for hypothesis rankingand their application for probably approximately correct (PAC)and expected loss (EL) learning criteria. Empirical results are provided to demonstrate the effectiveness of these ranking procedureson both synthetic datasets and real-world data from a spacecraft design optimization problem. 1 INTRODUCTION In many learning applications, the cost of information can be quite high, imposing a requirement that the learning algorithms glean as much usable information as possible with a minimum of data. For example: - In speedup learning, the expense of processing each training example can be significant [Tadepalli921. This paper provides a statistical decision-theoretic framework for the ranking of parametric distributions.