Grammar-Guided Evolutionary Construction of Bayesian networks

This paper proposes the EvoBANE system. EvoBANE automatically generates Bayesian networks for solving special-purpose problems. EvoBANE evolves a population of individuals that codify Bayesian networks until it finds near optimal individual that solves a given classification problem. EvoBANE has the flexibility to modify the constraints that condition the solution search space, self-adapting to the specifications of the problem to be solved. The system extends the GGEAS architecture. GGEAS is a general-purpose grammar-guided evolutionary automatic system, whose modular structure favors its application to the automatic construction of intelligent systems. EvoBANE has been applied to two classification benchmark datasets belonging to different application domains, and statistically compared with a genetic algorithm performing the same tasks. Results show that the proposed system performed better, as it manages different complexity constraints in order to find the simplest solution that best solves every problem.

Full scale test for explosion water barriers in small cross-section galleries

Underground coal mines explosions generally arise from the inflammation of a methane/air mixture. This explosion can also generate a subsequent coal dust explosion. Traditionally such explosions have being fought eliminating one or several of the factors needed by the explosion to take place. Although several preventive measures are taken to prevent explosions, other measures should be considered to reduce the effects or even to extinguish the flame front. Unlike other protection methods that remove one or two of the explosion triangle elements, namely; the ignition source, the oxidizing agent and the fuel, explosion barriers removes all of them: reduces the quantity of coal in suspension, cools the flame front and the steam generated by vaporization removes the oxygen present in the flame. Passive water barriers are autonomous protection systems against explosions that reduce to a satisfactory safety level the effects of methane and/or flammable dust explosions. The barriers are activated by the pressure wave provoked in the explosion destroying the barrier troughs and producing a uniform dispersion of the extinguishing agent throughout the gallery section in quantity enough to extinguish the explosion flame. Full scale tests have been carried out in Polish Barbara experimental mine at GIG Central Mining Institute in order to determine the requirements and the optimal installation conditions of these devices for small sections galleries which are very frequent in the Spanish coal mines. Full scale tests results have been analyzed to understand the explosion timing and development, in order to assess on the use of water barriers in the typical small crosssection Spanish galleries. Several arrangements of water barriers have been designed and tested to verify the effectiveness of the explosion suppression in each case. The results obtained demonstrate the efficiency of the water barriers in stopping the flame front even with smaller amounts of water than those established by the European standard. According to the tests realized, water barriers activation times are between 0.52 s and 0.78 s and the flame propagation speed are between 75 m/s and 80 m/s. The maximum pressures (Pmax) obtained in the full scale tests have varied between 0.2 bar and 1.8 bar. Passive barriers protect effectively against the spread of the flame but cannot be used as a safeguard of the gallery between the ignition source and the first row of water troughs or bags, or even after them, as the pressure could remain high after them even if the flame front has been extinguished.

The role of computational logic as a hinge paradigm among deduction, problem solving, programming, and parallelism

This paper presents some brief considerations on the role of Computational Logic in the construction of Artificial Intelligence systems and in programming in general. It does not address how the many problems in AI can be solved but, rather more modestly, tries to point out some advantages of Computational Logic as a tool for the AI scientist in his quest. It addresses the interaction between declarative and procedural views of programs (deduction and action), the impact of the intrinsic limitations of logic, the relationship with other apparently competing computational paradigms, and finally discusses implementation-related issues, such as the efficiency of current implementations and their capability for efficiently exploiting existing and future sequential and parallel hardware. The purpose of the discussion is in no way to present Computational Logic as the unique overall vehicle for the development of intelligent systems (in the firm belief that such a panacea is yet to be found) but rather to stress its strengths in providing reasonable solutions to several aspects of the task.

A Balanced Memory-Based Collaborative Filtering Similarity Measure.

Collaborative filtering recommender systems contribute to alleviating the problem of information overload that exists on the Internet as a result of the mass use of Web 2.0 applications. The use of an adequate similarity measure becomes a determining factor in the quality of the prediction and recommendation results of the recommender system, as well as in its performance. In this paper, we present a memory-based collaborative filtering similarity measure that provides extremely high-quality and balanced results; these results are complemented with a low processing time (high performance), similar to the one required to execute traditional similarity metrics. The experiments have been carried out on the MovieLens and Netflix databases, using a representative set of information retrieval quality measures.

A review on probabilistic graphical models in evolutionary computation

Thanks to their inherent properties, probabilistic graphical models are one of the prime candidates for machine learning and decision making tasks especially in uncertain domains. Their capabilities, like representation, inference and learning, if used effectively, can greatly help to build intelligent systems that are able to act accordingly in different problem domains. Evolutionary algorithms is one such discipline that has employed probabilistic graphical models to improve the search for optimal solutions in complex problems. This paper shows how probabilistic graphical models have been used in evolutionary algorithms to improve their performance in solving complex problems. Specifically, we give a survey of probabilistic model building-based evolutionary algorithms, called estimation of distribution algorithms, and compare different methods for probabilistic modeling in these algorithms.

Reliability and quality in the area of statistical thinking in Engineering

There is remarkable growing concern about the quality control at the time, which has led to the search for methods capable of addressing effectively the reliability analysis as part of the Statistic. Managers, researchers and Engineers must understand that 'statistical thinking' is not just a set of statistical tools. They should start considering 'statistical thinking' from a 'system', which means, developing systems that meet specific statistical tools and other methodologies for an activity. The aim of this article is to encourage them (engineers, researchers and managers) to develop a new way of thinking.

A fuzzy MCDA framework for safety assessment in the remediation of a uranium mill tailings site in Ukraine

The Pridneprovsky Chemical Plant was one of the largest uranium processing enterprises in the former USSR, producing a huge amount of uranium residues. The Zapadnoe tailings site contains most of these residues. We propose a theoretical framework based on multicriteria decision analysis and fuzzy logic to analyze different remediation alternatives for the Zapadnoe tailings, which simultaneously accounts for potentially conflicting economic, social and environmental objectives. We build an objective hierarchy that includes all the relevant aspects. Fuzzy rather than precise values are proposed for use to evaluate remediation alternatives against the different criteria and to quantify preferences, such as the weights representing the relative importance of criteria identified in the objective hierarchy. Finally, we suggest that remediation alternatives should be evaluated by means of a fuzzy additive multi-attribute utility function and ranked on the basis of the respective trapezoidal fuzzy number representing their overall utility.

An interactive method of fuzzy probability elicitation in risk analysis

Expert knowledge is used to assign probabilities to events in many risk analysis models. However, experts sometimes find it hard to provide specific values for these probabilities, preferring to express vague or imprecise terms that are mapped using a previously defined fuzzy number scale. The rigidity of these scales generates bias in the probability elicitation process and does not allow experts to adequately express their probabilistic judgments. We present an interactive method for extracting a fuzzy number from experts that represents their probabilistic judgments for a given event, along with a quality measure of the probabilistic judgments, useful in a final information filtering and analysis sensitivity process.

A Knowledge Model for Automatic Configuration of Traffic Messages

This paper describes a knowledge model for a configuration problem in the do-main of traffic control. The goal of this model is to help traffic engineers in the dynamic selection of a set of messages to be presented to drivers on variable message signals. This selection is done in a real-time context using data recorded by traffic detectors on motorways. The system follows an advanced knowledge-based solution that implements two abstract problem solving methods according to a model-based approach recently proposed in the knowledge engineering field. Finally, the paper presents a discussion about the advantages and drawbacks found for this problem as a consequence of the applied knowledge modeling ap-proach.

Using Electronic Documents for Knowledge Acquisition and Model Maintenance

Knowledge acquisition and model maintenance are key problems in knowledge engineering to improve the productivity in the development of intelligent systems. Although historically a number of technical solutions have been proposed in this area, the recent experience shows that there is still an important gap between the way end-users describe their expertise and the way intelligent systems represent knowledge. In this paper we propose an original way to cope with this problem based on electronic documents. We propose the concept of intelligent document processor as a tool that allows the end-user to read/write a document explaining how an intelligent system operates in such a way that, if the user changes the content of the document, the intelligent system will react to these changes. The paper presents the structure of such a document based on knowledge categories derived from the modern knowledge modeling methodologies together with a number of requirements to be understandable by end-users and problem solvers.

Watching subgraphs to improve efficiency in maximum clique search

This paper describes a new technique referred to as watched subgraphs which improves the performance of BBMC, a leading state of the art exact maximum clique solver (MCP). It is based on watched literals employed by modern SAT solvers for boolean constraint propagation. In efficient SAT algorithms, a list of clauses is kept for each literal (it is said that the clauses watch the literal) so that only those in the list are checked for constraint propagation when a (watched) literal is assigned during search. BBMC encodes vertex sets as bit strings, a bit block representing a subset of vertices (and the corresponding induced subgraph) the size of the CPU register word. The paper proposes to watch two subgraphs of critical sets during MCP search to efficiently compute a number of basic operations. Reported results validate the approach as the size and density of problem instances rise, while achieving comparable performance in the general case.

Smart Products: Una Revisión Sistemática de la Literatura

Existe una proliferación de los llamados Smart Products. Ello es debido a que cada vez se apueste más por este tipo de productos tanto en la vida cotidiana como en el sector industrial. Sin embargo el término Smart Product se utiliza con diferentes acepciones en diferentes contextos o dominios de aplicación. La utilización del término con una semántica diferente de la habitual en un contexto puede llevar a problemas serios de compresión. El objetivo de este trabajo es analizar las diferentes definiciones de Smart Products—Productos Inteligentes, Smart Products en terminología inglesa, ampliamente utilizada—que aparecen en la literatura con el objeto de estudiar los diferentes matices y alcances que ofrecen para valorar si es posible obtener una definición de consenso que satisfaga a todas las partes, y especificarla. Con el fin de poder abarcar definiciones conexas introducimos el concepto Smart Thing—este concepto incluirá aquellas definiciones que puedan estar relacionadas con los Smart Products, como es el caso de los Intelligent Products, Smart Objects, Intelligent Systems, Intelligent Object. Para poder analizar las diferentes definiciones existentes en la literatura existente realizamos una Revisión Sistemática de la Literatura. El enfoque de Computación Autonómica—Autonomic Computing—tiene varios aspectos en común con Smart Products. Por ello una vez analizadas las diferentes definiciones existentes en la literatura hemos procedido a estudiar los puntos en común que tienen con Autonomic Computing, con el fin de valorar si Autonomic Computing es un enfoque adecuado en el que nos podamos apoyar para especificar, y diseñar Smart Products.

Conditional density approximations with mixtures of polynomials

Mixtures of polynomials (MoPs) are a non-parametric density estimation technique especially designed for hybrid Bayesian networks with continuous and discrete variables. Algorithms to learn one- and multi-dimensional (marginal) MoPs from data have recently been proposed. In this paper we introduce two methods for learning MoP approximations of conditional densities from data. Both approaches are based on learning MoP approximations of the joint density and the marginal density of the conditioning variables, but they differ as to how the MoP approximation of the quotient of the two densities is found. We illustrate and study the methods using data sampled from known parametric distributions, and we demonstrate their applicability by learning models based on real neuroscience data. Finally, we compare the performance of the proposed methods with an approach for learning mixtures of truncated basis functions (MoTBFs). The empirical results show that the proposed methods generally yield models that are comparable to or significantly better than those found using the MoTBF-based method.

Identificación de los síntomas de daño o merma de capacidad resistente en estructuras en zonas sísmicas

Los sismos afectan a las estructuras en función de su intensidad. Normalmente se espera de las estructuras daños irreparables por motivos de ductilidad en los sismos nominales o de diseño, para protección de las personas y sus bienes. No obstante, las estructuras en zonas sísmicas sufren terremotos de baja o media intensidad de manera continuada y éstos pueden afectar a la capacidad resistente residual de las mismas, es por eso que en el presente trabajo se plantea lo siguiente: a) Identificar cuál es la estrategia o nivel de protección, que consideran las diferentes Normativas y Reglamentos frente a sismos de baja o mediana intensidad, puesto que durante la vida útil de una estructura, esta puede verse afectada por sismos de intensidad baja o moderada, los cuales también provocan daños; es por ello que es de mucha importancia conocer y estudiar el aporte, estrategias y demás parámetros que consideran las Normas, esto mediante la técnica de revisión de documentación o Literatura. b) Identificar la manera con que un terremoto de baja o media intensidad afecta a la capacidad resistente de las estructuras, sus señales, sus síntomas de daño, etc. Esto a través de tres técnicas de Investigación : Revisión en Literatura, Tormenta de ideas con un grupo de expertos en el tema, y mediante la Técnica Delphi; para finalmente aplicar una método de refinamiento para elaborar un listado y un mapa de síntomas esperables en las estructuras, consecuencia de eventos sísmicos de baja o mediana intensidad. Los cuales se podrían controlar con sistemas inteligentes y así monitorizar las construcciones para prever su comportamiento futuro. Earthquakes affect structures depending on its intensity. Normally it expected of the irreparable damage structures. It due to ductility in nominal earthquakes to protect people and property. Structures in seismic areas suffer earthquakes of low to medium intensity continually, and it may affect the residual resistant ability, therefore posed in this investigation is the following: (a) Identifying what is the strategy or level of protection, which consider different guidelines and regulations against earthquakes of low to medium intensity. Since during the service life of a structure may be affected by low or moderate intensity earthquakes, which also cause damage. For this reason it is very important also to meet and study the contribution, strategies and other parameters considered by the Guidelines by reviewing the documentation or literature technique. b) Identifying the way an earthquake of low to medium intensity affects the resistant ability of structures, their signs, their symptoms of injury, etc. Through three research techniques: review of documentation or literature, brainstorming technique with a group of experts, and using the Delphi technique. Finally applying a method of refining to produce a list and a map of symptoms expected in structures, consequence of low to medium intensity earthquakes. It could be controlled with intelligent systems and thus to monitor structures to predict its future behavior

MedVir: an interactive representation system of multidimensional medical data applied to Traumatic Brain Injury's rehabilitation prediction

Clinicians could model the brain injury of a patient through his brain activity. However, how this model is defined and how it changes when the patient is recovering are questions yet unanswered. In this paper, the use of MedVir framework is proposed with the aim of answering these questions. Based on complex data mining techniques, this provides not only the differentiation between TBI patients and control subjects (with a 72% of accuracy using 0.632 Bootstrap validation), but also the ability to detect whether a patient may recover or not, and all of that in a quick and easy way through a visualization technique which allows interaction.