The objective is to present one important aspect of the European IST-FET project "REV!GIS"1: the methodology which has been developed for the translation (interpretation) of the quality of the data into a "fitness for use" information, that we can confront to the user needs in its application. This methodology is based upon the notion of "ontologies" as a conceptual framework able to capture the explicit and implicit knowledge involved in the application. We do not address the general problem of formalizing such ontologies, instead, we rather try to illustrate this with three applications which are particular cases of the more general "data fusion" problem. In each application, we show how to deploy our methodology, by comparing several possible solutions, and we try to enlighten where are the quality issues, and what kind of solution to privilege, even at the expense of a highly complex computational approach. The expectation of the REV!GIS project is that computationally tractable solutions will be available among the next generation AI tools.

Data mining techniques on the biological analysis are spreading for most of the areas including the health care and medical information. We have applied the data mining techniques, such as KNN, SVM, MLP or decision trees over a unique dataset, which is collected from 16,380 analysis results for a year. Furthermore we have also used meta-classifiers to question the increased correlation rate between the liver disorder and the liver analysis outputs. The results show that there is a correlation among ALT, AST, Billirubin Direct and Billirubin Total down to 15% of error rate. Also the correlation coefficient is up to 94%. This makes possible to predict the analysis results from each other or disease patterns can be applied over the linear correlation of the parameters.

We reconsider the stochastic (sub)gradient approach to the unconstrained primal L1-SVM optimization. We observe that if the learning rate is inversely proportional to the number of steps, i.e., the number of times any training pattern is presented to the algorithm, the update rule may be transformed into the one of the classical perceptron with margin in which the margin threshold increases linearly with the number of steps. Moreover, if we cycle repeatedly through the possibly randomly permuted training set the dual variables defined naturally via the expansion of the weight vector as a linear combination of the patterns on which margin errors were made are shown to obey at the end of each complete cycle automatically the box constraints arising in dual optimization. This renders the dual Lagrangian a running lower bound on the primal objective tending to it at the optimum and makes available an upper bound on the relative accuracy achieved which provides a meaningful stopping criterion. In addition, we propose a mechanism of presenting the same pattern repeatedly to the algorithm which maintains the above properties. Finally, we give experimental evidence that algorithms constructed along these lines exhibit a considerably improved performance.

The Laplace approximation calls for the computation of second derivatives at the likelihood maximum. When the maximum is found by the EM algorithm, there is a convenient way to compute these derivatives. The likelihood gradient can be obtained from the EMauxiliary, while the Hessian can be obtained from this gradient with the Pearlmutter trick. Let X denote the observed data, H some hidden variables and Θ the model parameters. P (X, Θ) P (X, H, Θ) dH (2) has a more complex form.

The study of networks has received increased attention recently not only from the social sciences and statistics but also from physicists, computer scientists and mathematicians. One of the principal problem in networks is community detection. Many algorithms have been proposed for community finding but most of them do not have have theoretical guarantee for sparse networks and networks close to the phase transition boundary proposed by physicists. There are some exceptions but all have some incomplete theoretical basis. Here we propose an algorithm based on the graph distance of vertices in the network. We give theoretical guarantees that our method works in identifying communities for block models and can be extended for degree-corrected block models and block models with the number of communities growing with number of vertices. Despite favorable simulation results, we are not yet able to conclude that our method is satisfactory for worst possible case. We illustrate on a network of political blogs, Facebook networks and some other networks.

Generative Stochastic Networks (GSNs) have been recently introduced as an alternative to traditional probabilistic modeling: instead of parametrizing the data distribution directly, one parametrizes a transition operator for a Markov chain whose stationary distribution is an estimator of the data generating distribution. The result of training is therefore a machine that generates samples through this Markov chain. However, the previously introduced GSN consistency theorems suggest that in order to capture a wide class of distributions, the transition operator in general should be multimodal, something that has not been done before this paper. We introduce for the first time multimodal transition distributions for GSNs, in particular using models in the NADE family (Neural Autoregressive Density Estimator) as output distributions of the transition operator. A NADE model is related to an RBM (and can thus model multimodal distributions) but its likelihood (and likelihood gradient) can be computed easily. The parameters of the NADE are obtained as a learned function of the previous state of the learned Markov chain. Experiments clearly illustrate the advantage of such multimodal transition distributions over unimodal GSNs.

Distributional semantic models derive computational representations of word meaning from the patterns of co-occurrence of words in text. Such models have been a success story of computational linguistics, being able to provide reliable estimates of semantic relatedness for the many semantic tasks requiring them. However, distributional models extract meaning information exclusively from text, which is an extremely impoverished basis compared to the rich perceptual sources that ground human semantic knowledge. We address the lack of perceptual grounding of distributional models by exploiting computer vision techniques that automatically identify discrete visual words in images, so that the distributional representation of a word can be extended to also encompass its co-occurrence with the visual words of images it is associated with. We propose a flexible architecture to integrate text- and image-based distributional information, and we show in a set of empirical tests that our integrated model is superior to the purely text-based approach, and it provides somewhat complementary semantic information with respect to the latter.

Motivated by an application in computational biology, we consider low-rank matrix factorization with $\{0,1\}$-constraints on one of the factors and optionally convex constraints on the second one. In addition to the non-convexity shared with other matrix factorization schemes, our problem is further complicated by a combinatorial constraint set of size $2^{m \cdot r}$, where $m$ is the dimension of the data points and $r$ the rank of the factorization. Despite apparent intractability, we provide - in the line of recent work on non-negative matrix factorization by Arora et al. (2012) - an algorithm that provably recovers the underlying factorization in the exact case with $O(m r 2^r + mnr + r^2 n)$ operations for $n$ datapoints. To obtain this result, we use theory around the Littlewood-Offord lemma from combinatorics.

Hierarchical parametric models consisting of observable and latent variables are widely used for unsupervised learning tasks. For example, a mixture model is a representative hierarchical model for clustering. From the statistical point of view, the models can be regular or singular due to the distribution of data. In the regular case, the models have the identifiability; there is one-to-one relation between a probability density function for the model expression and the parameter. The Fisher information matrix is positive definite, and the estimation accuracy of both observable and latent variables has been studied. In the singular case, on the other hand, the models are not identifiable and the Fisher matrix is not positive definite. Conventional statistical analysis based on the inverse Fisher matrix is not applicable. Recently, an algebraic geometrical analysis has been developed and is used to elucidate the Bayes estimation of observable variables. The present paper applies this analysis to latent-variable estimation and determines its theoretical performance. Our results clarify behavior of the convergence of the posterior distribution. It is found that the posterior of the observable-variable estimation can be different from the one in the latent-variable estimation. Because of the difference, the Markov chain Monte Carlo method based on the parameter and the latent variable cannot construct the desired posterior distribution.

We introduce a general framework for measuring risk in the context of Markov control processes with risk maps on general Borel spaces that generalize known concepts of risk measures in mathematical finance, operations research and behavioral economics. Within the framework, applying weighted norm spaces to incorporate also unbounded costs, we study two types of infinite-horizon risk-sensitive criteria, discounted total risk and average risk, and solve the associated optimization problems by dynamic programming. For the discounted case, we propose a new discount scheme, which is different from the conventional form but consistent with the existing literature, while for the average risk criterion, we state Lyapunov-like stability conditions that generalize known conditions for Markov chains to ensure the existence of solutions to the optimality equation.