Standard models of multi-agent modal logic do not capture the fact that information is often ambiguous, and may be interpreted in different ways by different agents. We propose a framework that can model this, and consider different semantics that capture different assumptions about the agents' beliefs regarding whether or not there is ambiguity. We consider the impact of ambiguity on a seminal result in economics: Aumann's result saying that agents with a common prior cannot agree to disagree. This result is known not to hold if agents do not have a common prior; we show that it also does not hold in the presence of ambiguity. We then consider the tradeoff between assuming a common interpretation (i.e., no ambiguity) and a common prior (i.e., shared initial beliefs).

The study is from a base of accident scenarii in rail transport (feedback) in order to develop a tool to share build and sustain knowledge and safety and secondly to exploit the knowledge stored to prevent the reproduction of accidents / incidents. This tool should ultimately lead to the proposal of prevention and protection measures to minimize the risk level of a new transport system and thus to improve safety. The approach to achieving this goal largely depends on the use of artificial intelligence techniques and rarely the use of a method of automatic learning in order to develop a feasibility model of a software tool based on case based reasoning (CBR) to exploit stored knowledge in order to create know-how that can help stimulate domain experts in the task of analysis, evaluation and certification of a new system. Index Terms-- Accident scenario, Exploitation of knowledge Return of experience, Case based reasoning,Security.

In the paper, frameworks for electronic shopping of composite (modular) products are described: (a) multicriteria selection (product is considered as a whole system, it is a traditional approach), (b) combinatorial synthesis (composition) of the product from its components, (c) aggregation of the product from several selected products/prototypes. The following product model is examined: (i) general tree-like structure, (ii) set of system parts/components (leaf nodes), (iii) design alternatives (DAs) for each component, (iv) ordinal priorities for DAs, and (v) estimates of compatibility between DAs for different components. The combinatorial synthesis is realized as morphological design of a composite (modular) product or an extended composite product (e.g., product and support services as financial instruments). Here the solving process is based on Hierarchical Morphological Multicriteria Design (HMMD): (i) multicriteria selection of alternatives for system parts, (ii) composing the selected alternatives into a resultant combination (while taking into account ordinal quality of the alternatives above and their compatibility). The aggregation framework is based on consideration of aggregation procedures, for example: (i) addition procedure: design of a products substructure or an extended substructure ('kernel') and addition of elements, and (ii) design procedure: design of the composite solution based on all elements of product superstructure. Applied numerical examples (e.g., composite product, extended composite product, product repair plan, and product trajectory) illustrate the proposed approaches.

We define and discuss the first sparse coding algorithm based on closed-form EM updates and continuous latent variables. The underlying generative model consists of a standard `spike-and-slab' prior and a Gaussian noise model. Closed-form solutions for E- and M-step equations are derived by generalizing probabilistic PCA. The resulting EM algorithm can take all modes of a potentially multi-modal posterior into account. The computational cost of the algorithm scales exponentially with the number of hidden dimensions. However, with current computational resources, it is still possible to efficiently learn model parameters for medium-scale problems. Thus the model can be applied to the typical range of source separation tasks. In numerical experiments on artificial data we verify likelihood maximization and show that the derived algorithm recovers the sparse directions of standard sparse coding distributions. On source separation benchmarks comprised of realistic data we show that the algorithm is competitive with other recent methods.

Language evolution might have preferred certain prior social configurations over others. Experiments conducted with models of different social structures (varying subgroup interactions and the role of a dominant interlocutor) suggest that having isolated agent groups rather than an interconnected agent is more advantageous for the emergence of a social communication system. Distinctive groups that are closely connected by communication yield systems less like natural language than fully isolated groups inhabiting the same world. Furthermore, the addition of a dominant male who is asymmetrically favoured as a hearer, and equally likely to be a speaker has no positive influence on the disjoint groups.

Sparsity-promoting priors have become increasingly popular over recent years due to an increased number of regression and classification applications involving a large number of predictors. In time series applications where observations are collected over time, it is often unrealistic to assume that the underlying sparsity pattern is fixed. We propose here an original class of flexible Bayesian linear models for dynamic sparsity modelling. The proposed class of models expands upon the existing Bayesian literature on sparse regression using generalized multivariate hyperbolic distributions. The properties of the models are explored through both analytic results and simulation studies. We demonstrate the model on a financial application where it is shown that it accurately represents the patterns seen in the analysis of stock and derivative data, and is able to detect major events by filtering an artificial portfolio of assets.

Personalized web services strive to adapt their services (advertisements, news articles, etc) to individual users by making use of both content and user information. Despite a few recent advances, this problem remains challenging for at least two reasons. First, web service is featured with dynamically changing pools of content, rendering traditional collaborative filtering methods inapplicable. Second, the scale of most web services of practical interest calls for solutions that are both fast in learning and computation. In this work, we model personalized recommendation of news articles as a contextual bandit problem, a principled approach in which a learning algorithm sequentially selects articles to serve users based on contextual information about the users and articles, while simultaneously adapting its article-selection strategy based on user-click feedback to maximize total user clicks. The contributions of this work are three-fold. First, we propose a new, general contextual bandit algorithm that is computationally efficient and well motivated from learning theory. Second, we argue that any bandit algorithm can be reliably evaluated offline using previously recorded random traffic. Finally, using this offline evaluation method, we successfully applied our new algorithm to a Yahoo! Front Page Today Module dataset containing over 33 million events. Results showed a 12.5% click lift compared to a standard context-free bandit algorithm, and the advantage becomes even greater when data gets more scarce.

We introduce here mlpy, a library providing access to a wide spectrum of machine learning methods implemented in Python, which has proven to be an effective environment for building scientific oriented tools (Pérez et al., 2011). Although planned for general purpose applications, mlpy has the computational biology in general, and the functional genomics modeling in particular, as the elective application fields. As a major applications example, we use mlpy methods to implement molecular profiling experiments that need to warrant study reproducibility(Ioannidis et al., 2009) and flawless results(Ambroise and McLachlan, 2002). This task requires the availability of highly modular tools allowing the practioners to build an adequate workflow for the task at hand following authoritative guidelines (The MicroArray Quality Control (MAQC) Consortium, 2010). Such workflow involves a complex sequence of steps, both in the development and in the validation phases, starting from the upstream preprocessing algorithms to the downstream predictive analysis, repeated several times to accommodate the resampling schema. The dimension of highthroughtput data involved (thousands of samples described by millions of features) and the large number of replicates needed to control bias effects make also efficiency an essential requirement.

Risk bounds for Classification and Regression Trees (CART, Breiman et. al. 1984) classifiers are obtained under a margin condition in the binary supervised classification framework. These risk bounds are obtained conditionally on the construction of the maximal deep binary tree and permit to prove that the linear penalty used in the CART pruning algorithm is valid under a margin condition. It is also shown that, conditionally on the construction of the maximal tree, the final selection by test sample does not alter dramatically the estimation accuracy of the Bayes classifier. In the two-class classification framework, the risk bounds that are proved, obtained by using penalized model selection, validate the CART algorithm which is used in many data mining applications such as Biology, Medicine or Image Coding.

Contextual bandit algorithms have become popular for online recommendation systems such as Digg, Yahoo! Buzz, and news recommendation in general. \emph{Offline} evaluation of the effectiveness of new algorithms in these applications is critical for protecting online user experiences but very challenging due to their "partial-label" nature. Common practice is to create a simulator which simulates the online environment for the problem at hand and then run an algorithm against this simulator. However, creating simulator itself is often difficult and modeling bias is usually unavoidably introduced. In this paper, we introduce a \emph{replay} methodology for contextual bandit algorithm evaluation. Different from simulator-based approaches, our method is completely data-driven and very easy to adapt to different applications. More importantly, our method can provide provably unbiased evaluations. Our empirical results on a large-scale news article recommendation dataset collected from Yahoo! Front Page conform well with our theoretical results. Furthermore, comparisons between our offline replay and online bucket evaluation of several contextual bandit algorithms show accuracy and effectiveness of our offline evaluation method.