This paper proposes a practical optimization method for layered neural networks, by which the optimal model and parameter can be found simultaneously. 'i\Te modify the conventional information criterion into a differentiable function of parameters, and then, minimize it,while controlling it back to the ordinary form. Effectiveness of this method is discussed theoretically and experimentally.

We propose a method for improving the performance of any network designedto predict the next value of a time series. Vve advocate analyzing the deviations of the network's predictions from the data in the training set. This can be carried out by a secondary network trainedon the time series of these residuals. The combined system of the two networks is viewed as the new predictor. We demonstrate the simplicity and success of this method, by applying itto the sunspots data. The small corrections of the secondary network can be regarded as resulting from a Taylor expansion of a complex network which includes the combined system.

In many cases the crosscorrelations betweenthe activities of cortical neurons are approximately symmetric about zero time delay. These have been taken as an indication of the presence of "functional connectivity" between the correlated neurons (Fetz, Toyama and Smith 1991, Abeles 1991). However, a quantitative comparison between the observed cross-correlations and those expected to exist between neurons that are part of a large assembly of interacting population has been lacking. Most of the theoretical studies of recurrent neural network models consider only time averaged firing rates, which are usually given as solutions of mean-field equations. They do not account for the fluctuations about these averages, the study of which requires going beyond the mean-field approximations. In this work we perform a theoretical study of the fluctuations in the neuronal activities and their correlations, in a large stochastic network of excitatory and inhibitory neurons. Depending on the model parameters, this system can exhibit coherent undamped oscillations. Here we focus on parameter regimes where the system is in a statistically stationary state, which is more appropriate for modeling non oscillatory neuronal activity in cortex. Our results for the magnitudes and the time-dependence of the correlation functions can provide a basis for comparison with physiological data on neuronal correlation functions.

The vast amounts of on-line text now available have ledto renewed interest in information extraction (IE) systems thatanalyze unrestricted text, producing a structured representation ofselected information from the text. This paper presents a novel approachthat uses machine learning to acquire knowledge for some of the higher level IE processing. Wrap-Up is a trainable IE discourse component that makes intersentential inferences and identifies logicalrelations among information extracted from the text. Previous corpus-based approaches were limited to lower level processing such as part-of-speech tagging, lexical disambiguation, and dictionary construction. Wrap-Up is fully trainable, and not onlyautomatically decides what classifiers are needed, but even derives the featureset for each classifier automatically. Performance equals that of a partially trainable discourse module requiring manual customization for each domain.

This paper is a multidisciplinary review of empirical, statistical learning from a graphical model perspective. Well-known examples of graphical models include Bayesian networks, directed graphs representing a Markov chain, and undirected networks representing a Markov field. These graphical models are extended to model data analysis and empirical learning using the notation of plates. Graphical operations for simplifying and manipulating a problem are provided including decomposition, differentiation, andthe manipulation of probability models from the exponential family. Two standard algorithm schemas for learning are reviewed in a graphical framework: Gibbs sampling and the expectation maximizationalgorithm. Using these operations and schemas, some popular algorithms can be synthesized from their graphical specification. This includes versions of linear regression, techniques for feed-forward networks, and learning Gaussian and discrete Bayesian networks from data. The paper concludes by sketching some implications for data analysis and summarizing how some popular algorithms fall within the framework presented. The main original contributions here are the decompositiontechniques and the demonstration that graphical models provide a framework for understanding and developing complex learning algorithms.

Information extraction is the task of automaticallypicking up information of interest from an unconstrained text. Informationof interest is usually extracted in two steps. First, sentence level processing locates relevant pieces of information scatteredthroughout the text; second, discourse processing merges coreferential information to generate the output. In the first step, pieces of information are locally identified without recognizing any relationships among them. A key word search or simple patternsearch can achieve this purpose. The second step requires deeperknowledge in order to understand relationships among separately identified pieces of information. Previous information extraction systems focused on the first step, partly because they were not required to link up each piece of information with other pieces. To link the extracted pieces of information and map them onto a structuredoutput format, complex discourse processing is essential. This paperreports on a Japanese information extraction system that merges information using a pattern matcher and discourse processor. Evaluationresults show a high level of system performance which approaches human performance.

This article summarizes the findings of a 1992 study of knowledge-based systems research and applications in Japan. Representatives of universities and businesses were chosen by the Japan Technology Evaluation Center to investigate the state of the technology in Japan relative to the United States. The panel's report focused on applications, tools, and research and development in universities and industry and on major national projects.

The Hong Kong-based Mass Transit Railway Corp. (MTRC) has developed the Station Management Expert e Norwegian Police Data Center help predict aircraft fires and other System (SMES). SMES is an intelligent utilized an expert system to catastrophes. The police put and risk factors from the records functions and advising the controller the intelligent application online to of the National Transportation Safety of actions to take in case of emergency. The system is installed in Ya Ma at the games while complying with Carnegie Group and Westinghouse Tei Station as a test site, and the complex national employment regulations. Electric (both in Pittsburgh, Penn.) are MTRC plans to expand its use Plans are to deploy and network working with Pittsburgh area medical throughout the subway system as it the expert system into every law centers to develop an intelligent proves to be successful. The network Martin Marietta (Bethesda, Md.) is developed a neural network application will gather and organize data on using a real-time expert system to that has improved the efficiency clinical diagnoses, treatment, clinical build the Traffic Operations Center of its direct mail marketing efforts by and research findings, and patient (TOC) component of its Intelligent 35%.

The theory revision problem is the problem of how best to go about revising a deficient domain theory using information contained in examples that expose inaccuracies. In this paper we present our approach to the theory revision problem for propositional domain theories. The approach described here, called PTR, uses probabilities associated with domain theory elements to numerically track the ``flow'' of proof through the theory. This allows us to measure the precise role of a clause or literal in allowing or preventing a (desired or undesired) derivation for a given example. This information is used to efficiently locate and repair flawed elements of the theory. PTR is proved to converge to a theory which correctly classifies all examples, and shown experimentally to be fast and accurate even for deep theories.

Typical methods for gradient descent in neural network learning involve calculation of derivatives based on a detailed knowledge of the network model. This requires extensive, time consuming calculations for each pattern presentation and high precision that makes it difficult to implement in VLSI. We present here a perturbation technique that measures, not calculates, the gradient. Since the technique uses the actual network as a measuring device, errors in modeling neuron activation and synaptic weights do not cause errors in gradient descent. The method is parallel in nature and easy to implement in VLSI. We describe the theory of such an algorithm, an analysis of its domain of applicability, some simulations using it and an outline of a hardware implementation.