Modelado de sistemas multimedia para personalización y recomendación híbrida a partir del consumo audiovisual de los usuarios

This doctoral thesis focuses on the modeling of multimedia systems to create personalized recommendation services based on the analysis of users’ audiovisual consumption. Research is focused on the characterization of both users’ audiovisual consumption and content, specifically images and video. This double characterization converges into a hybrid recommendation algorithm, adapted to different application scenarios covering different specificities and constraints. Hybrid recommendation systems use both content and user information as input data, applying the knowledge from the analysis of these data as the initial step to feed the algorithms in order to generate personalized recommendations. Regarding the user information, this doctoral thesis focuses on the analysis of audiovisual consumption to infer implicitly acquired preferences. The inference process is based on a new probabilistic model proposed in the text. This model takes into account qualitative and quantitative consumption factors on the one hand, and external factors such as zapping factor or company factor on the other. As for content information, this research focuses on the modeling of descriptors and aesthetic characteristics, which influence the user and are thus useful for the recommendation system. Similarly, the automatic extraction of these descriptors from the audiovisual piece without excessive computational cost has been considered a priority, in order to ensure applicability to different real scenarios. Finally, a new content-based recommendation algorithm has been created from the previously acquired information, i.e. user preferences and content descriptors. This algorithm has been hybridized with a collaborative filtering algorithm obtained from the current state of the art, so as to compare the efficiency of this hybrid recommender with the individual techniques of recommendation (different hybridization techniques of the state of the art have been studied for suitability). The content-based recommendation focuses on the influence of the aesthetic characteristics on the users. The heterogeneity of the possible users of these kinds of systems calls for the use of different criteria and attributes to create effective recommendations. Therefore, the proposed algorithm is adaptable to different perceptions producing a dynamic representation of preferences to obtain personalized recommendations for each user of the system. The hypotheses of this doctoral thesis have been validated by conducting a set of tests with real users, or by querying a database containing user preferences - available to the scientific community. This thesis is structured based on the different research and validation methodologies of the techniques involved. In the three central chapters the state of the art is studied and the developed algorithms and models are validated via self-designed tests. It should be noted that some of these tests are incremental and confirm the validation of previously discussed techniques. Resumen Esta tesis doctoral se centra en el modelado de sistemas multimedia para la creación de servicios personalizados de recomendación a partir del análisis de la actividad de consumo audiovisual de los usuarios. La investigación se focaliza en la caracterización tanto del consumo audiovisual del usuario como de la naturaleza de los contenidos, concretamente imágenes y vídeos. Esta doble caracterización de usuarios y contenidos confluye en un algoritmo de recomendación híbrido que se adapta a distintos escenarios de aplicación, cada uno de ellos con distintas peculiaridades y restricciones. Todo sistema de recomendación híbrido toma como datos de partida tanto información del usuario como del contenido, y utiliza este conocimiento como entrada para algoritmos que permiten generar recomendaciones personalizadas. Por la parte de la información del usuario, la tesis se centra en el análisis del consumo audiovisual para inferir preferencias que, por lo tanto, se adquieren de manera implícita. Para ello, se ha propuesto un nuevo modelo probabilístico que tiene en cuenta factores de consumo tanto cuantitativos como cualitativos, así como otros factores de contorno, como el factor de zapping o el factor de compañía, que condicionan la incertidumbre de la inferencia. En cuanto a la información del contenido, la investigación se ha centrado en la definición de descriptores de carácter estético y morfológico que resultan influyentes en el usuario y que, por lo tanto, son útiles para la recomendación. Del mismo modo, se ha considerado una prioridad que estos descriptores se puedan extraer automáticamente de un contenido sin exigir grandes requisitos computacionales y, de tal forma que se garantice la posibilidad de aplicación a escenarios reales de diverso tipo. Por último, explotando la información de preferencias del usuario y de descripción de los contenidos ya obtenida, se ha creado un nuevo algoritmo de recomendación basado en contenido. Este algoritmo se cruza con un algoritmo de filtrado colaborativo de referencia en el estado del arte, de tal manera que se compara la eficiencia de este recomendador híbrido (donde se ha investigado la idoneidad de las diferentes técnicas de hibridación del estado del arte) con cada una de las técnicas individuales de recomendación. El algoritmo de recomendación basado en contenido que se ha creado se centra en las posibilidades de la influencia de factores estéticos en los usuarios, teniendo en cuenta que la heterogeneidad del conjunto de usuarios provoca que los criterios y atributos que condicionan las preferencias de cada individuo sean diferentes. Por lo tanto, el algoritmo se adapta a las diferentes percepciones y articula una metodología dinámica de representación de las preferencias que permite obtener recomendaciones personalizadas, únicas para cada usuario del sistema. Todas las hipótesis de la tesis han sido debidamente validadas mediante la realización de pruebas con usuarios reales o con bases de datos de preferencias de usuarios que están a disposición de la comunidad científica. La diferente metodología de investigación y validación de cada una de las técnicas abordadas condiciona la estructura de la tesis, de tal manera que los tres capítulos centrales se estructuran sobre su propio estudio del estado del arte y los algoritmos y modelos desarrollados se validan mediante pruebas autónomas, sin impedir que, en algún caso, las pruebas sean incrementales y ratifiquen la validación de técnicas expuestas anteriormente.

A model for turbulent combustion simulation of large scale hydrogen explosions

El agotamiento, la ausencia o, simplemente, la incertidumbre sobre la cantidad de las reservas de combustibles fósiles se añaden a la variabilidad de los precios y a la creciente inestabilidad en la cadena de aprovisionamiento para crear fuertes incentivos para el desarrollo de fuentes y vectores energéticos alternativos. El atractivo de hidrógeno como vector energético es muy alto en un contexto que abarca, además, fuertes inquietudes por parte de la población sobre la contaminación y las emisiones de gases de efecto invernadero. Debido a su excelente impacto ambiental, la aceptación pública del nuevo vector energético dependería, a priori, del control de los riesgos asociados su manipulación y almacenamiento. Entre estos, la existencia de un innegable riesgo de explosión aparece como el principal inconveniente de este combustible alternativo. Esta tesis investiga la modelización numérica de explosiones en grandes volúmenes, centrándose en la simulación de la combustión turbulenta en grandes dominios de cálculo en los que la resolución que es alcanzable está fuertemente limitada. En la introducción, se aborda una descripción general de los procesos de explosión. Se concluye que las restricciones en la resolución de los cálculos hacen necesario el modelado de los procesos de turbulencia y de combustión. Posteriormente, se realiza una revisión crítica de las metodologías disponibles tanto para turbulencia como para combustión, que se lleva a cabo señalando las fortalezas, deficiencias e idoneidad de cada una de las metodologías. Como conclusión de esta investigación, se obtiene que la única estrategia viable para el modelado de la combustión, teniendo en cuenta las limitaciones existentes, es la utilización de una expresión que describa la velocidad de combustión turbulenta en función de distintos parámetros. Este tipo de modelos se denominan Modelos de velocidad de llama turbulenta y permiten cerrar una ecuación de balance para la variable de progreso de combustión. Como conclusión también se ha obtenido, que la solución más adecuada para la simulación de la turbulencia es la utilización de diferentes metodologías para la simulación de la turbulencia, LES o RANS, en función de la geometría y de las restricciones en la resolución de cada problema particular. Sobre la base de estos hallazgos, el crea de un modelo de combustión en el marco de los modelos de velocidad de la llama turbulenta. La metodología propuesta es capaz de superar las deficiencias existentes en los modelos disponibles para aquellos problemas en los que se precisa realizar cálculos con una resolución moderada o baja. Particularmente, el modelo utiliza un algoritmo heurístico para impedir el crecimiento del espesor de la llama, una deficiencia que lastraba el célebre modelo de Zimont. Bajo este enfoque, el énfasis del análisis se centra en la determinación de la velocidad de combustión, tanto laminar como turbulenta. La velocidad de combustión laminar se determina a través de una nueva formulación capaz de tener en cuenta la influencia simultánea en la velocidad de combustión laminar de la relación de equivalencia, la temperatura, la presión y la dilución con vapor de agua. La formulación obtenida es válida para un dominio de temperaturas, presiones y dilución con vapor de agua más extenso de cualquiera de las formulaciones previamente disponibles. Por otra parte, el cálculo de la velocidad de combustión turbulenta puede ser abordado mediante el uso de correlaciones que permiten el la determinación de esta magnitud en función de distintos parámetros. Con el objetivo de seleccionar la formulación más adecuada, se ha realizado una comparación entre los resultados obtenidos con diversas expresiones y los resultados obtenidos en los experimentos. Se concluye que la ecuación debida a Schmidt es la más adecuada teniendo en cuenta las condiciones del estudio. A continuación, se analiza la importancia de las inestabilidades de la llama en la propagación de los frentes de combustión. Su relevancia resulta significativa para mezclas pobres en combustible en las que la intensidad de la turbulencia permanece moderada. Estas condiciones son importantes dado que son habituales en los accidentes que ocurren en las centrales nucleares. Por ello, se lleva a cabo la creación de un modelo que permita estimar el efecto de las inestabilidades, y en concreto de la inestabilidad acústica-paramétrica, en la velocidad de propagación de llama. El modelado incluye la derivación matemática de la formulación heurística de Bauwebs et al. para el cálculo de la incremento de la velocidad de combustión debido a las inestabilidades de la llama, así como el análisis de la estabilidad de las llamas con respecto a una perturbación cíclica. Por último, los resultados se combinan para concluir el modelado de la inestabilidad acústica-paramétrica. Tras finalizar esta fase, la investigación se centro en la aplicación del modelo desarrollado en varios problemas de importancia para la seguridad industrial y el posterior análisis de los resultados y la comparación de los mismos con los datos experimentales correspondientes. Concretamente, se abordo la simulación de explosiones en túneles y en contenedores, con y sin gradiente de concentración y ventilación. Como resultados generales, se logra validar el modelo confirmando su idoneidad para estos problemas. Como última tarea, se ha realizado un analisis en profundidad de la catástrofe de Fukushima-Daiichi. El objetivo del análisis es determinar la cantidad de hidrógeno que explotó en el reactor número uno, en contraste con los otros estudios sobre el tema que se han centrado en la determinación de la cantidad de hidrógeno generado durante el accidente. Como resultado de la investigación, se determinó que la cantidad más probable de hidrogeno que fue consumida durante la explosión fue de 130 kg. Es un hecho notable el que la combustión de una relativamente pequeña cantidad de hidrogeno pueda causar un daño tan significativo. Esta es una muestra de la importancia de este tipo de investigaciones. Las ramas de la industria para las que el modelo desarrollado será de interés abarca la totalidad de la futura economía de hidrógeno (pilas de combustible, vehículos, almacenamiento energético, etc) con un impacto especial en los sectores del transporte y la energía nuclear, tanto para las tecnologías de fisión y fusión. ABSTRACT The exhaustion, absolute absence or simply the uncertainty on the amount of the reserves of fossil fuels sources added to the variability of their prices and the increasing instability and difficulties on the supply chain are strong incentives for the development of alternative energy sources and carriers. The attractiveness of hydrogen in a context that additionally comprehends concerns on pollution and emissions is very high. Due to its excellent environmental impact, the public acceptance of the new energetic vector will depend on the risk associated to its handling and storage. Fromthese, the danger of a severe explosion appears as the major drawback of this alternative fuel. This thesis investigates the numerical modeling of large scale explosions, focusing on the simulation of turbulent combustion in large domains where the resolution achievable is forcefully limited. In the introduction, a general description of explosion process is undertaken. It is concluded that the restrictions of resolution makes necessary the modeling of the turbulence and combustion processes. Subsequently, a critical review of the available methodologies for both turbulence and combustion is carried out pointing out their strengths and deficiencies. As a conclusion of this investigation, it appears clear that the only viable methodology for combustion modeling is the utilization of an expression for the turbulent burning velocity to close a balance equation for the combustion progress variable, a model of the Turbulent flame velocity kind. Also, that depending on the particular resolution restriction of each problem and on its geometry the utilization of different simulation methodologies, LES or RANS, is the most adequate solution for modeling the turbulence. Based on these findings, the candidate undertakes the creation of a combustion model in the framework of turbulent flame speed methodology which is able to overcome the deficiencies of the available ones for low resolution problems. Particularly, the model utilizes a heuristic algorithm to maintain the thickness of the flame brush under control, a serious deficiency of the Zimont model. Under the approach utilized by the candidate, the emphasis of the analysis lays on the accurate determination of the burning velocity, both laminar and turbulent. On one side, the laminar burning velocity is determined through a newly developed correlation which is able to describe the simultaneous influence of the equivalence ratio, temperature, steam dilution and pressure on the laminar burning velocity. The formulation obtained is valid for a larger domain of temperature, steam dilution and pressure than any of the previously available formulations. On the other side, a certain number of turbulent burning velocity correlations are available in the literature. For the selection of the most suitable, they have been compared with experiments and ranked, with the outcome that the formulation due to Schmidt was the most adequate for the conditions studied. Subsequently, the role of the flame instabilities on the development of explosions is assessed. Their significance appears to be of importance for lean mixtures in which the turbulence intensity remains moderate. These are important conditions which are typical for accidents on Nuclear Power Plants. Therefore, the creation of a model to account for the instabilities, and concretely, the acoustic parametric instability is undertaken. This encloses the mathematical derivation of the heuristic formulation of Bauwebs et al. for the calculation of the burning velocity enhancement due to flame instabilities as well as the analysis of the stability of flames with respect to a cyclic velocity perturbation. The results are combined to build a model of the acoustic-parametric instability. The following task in this research has been to apply the model developed to several problems significant for the industrial safety and the subsequent analysis of the results and comparison with the corresponding experimental data was performed. As a part of such task simulations of explosions in a tunnel and explosions in large containers, with and without gradient of concentration and venting have been carried out. As a general outcome, the validation of the model is achieved, confirming its suitability for the problems addressed. As a last and final undertaking, a thorough study of the Fukushima-Daiichi catastrophe has been carried out. The analysis performed aims at the determination of the amount of hydrogen participating on the explosion that happened in the reactor one, in contrast with other analysis centered on the amount of hydrogen generated during the accident. As an outcome of the research, it was determined that the most probable amount of hydrogen exploding during the catastrophe was 130 kg. It is remarkable that the combustion of such a small quantity of material can cause tremendous damage. This is an indication of the importance of these types of investigations. The industrial branches that can benefit from the applications of the model developed in this thesis include the whole future hydrogen economy, as well as nuclear safety both in fusion and fission technology.

An Efficient Hybrid GO-PWS Algorithm to Analyze Conformal Serrated-Edge Reflectors for Millimeter-Wave Compact Range

A method to analyze parabolic reflectors with arbitrary piecewise rim is presented in this communication. This kind of reflectors, when operating as collimators in compact range facilities, needs to be large in terms of wavelength. Their analysis is very inefficient, when it is carried out with fullwave/MoM techniques, and it is not very appropriate for designing with PO techniques. Also, fast GO formulations do not offer enough accuracy to reach performance results. The proposed algorithm is based on a GO-PWS hybrid scheme, using analytical as well as non-analytical formulations. On one side, an analytical treatment of the polygonal rim reflectors is carried out. On the other side, non-analytical calculi are based on efficient operations, such as M2 order 2-dimensional FFT. A combination of these two techniques in the algorithm ensures real ad-hoc design capabilities, reached through analysis speedup. The purpose of the algorithm is to obtain an optimal conformal serrated-edge reflector design through the analysis of the field quality within the quiet zone that it is able to generate in its forward half space.

Inverse kinematics of a humanoid robot with non-spherical hip: a hybrid algorithm approach

This paper describes an approach to solve the inverse kinematics problem of humanoid robots whose construction shows a small but non negligible offset at the hip which prevents any purely analytical solution to be developed. Knowing that a purely numerical solution is not feasible due to variable efficiency problems, the proposed one first neglects the offset presence in order to obtain an approximate “solution” by means of an analytical algorithm based on screw theory, and then uses it as the initial condition of a numerical refining procedure based on the Levenberg‐Marquardt algorithm. In this way, few iterations are needed for any specified attitude, making it possible to implement the algorithm for real‐time applications. As a way to show the algorithm’s implementation, one case of study is considered throughout the paper, represented by the SILO2 humanoid robot.

Hybrid Monte Carlo dose algorithm for low energy X-rays intraoperative radiation therapy

Low energy X-rays Intra-Operative Radiation Therapy (XIORT) treatment delivered during surgery (ex: INTRABEAM, Carl Zeiss, and Axxent, Xoft) can benefit from accurate and fast dose prediction in a patient 3D volume.

Hybrid Approach Combining Machine Learning and a Rule-Based Expert System for Text Categorization

This paper discusses a novel hybrid approach for text categorization that combines a machine learning algorithm, which provides a base model trained with a labeled corpus, with a rule-based expert system, which is used to improve the results provided by the previous classifier, by filtering false positives and dealing with false negatives. The main advantage is that the system can be easily fine-tuned by adding specific rules for those noisy or conflicting categories that have not been successfully trained. We also describe an implementation based on k-Nearest Neighbor and a simple rule language to express lists of positive, negative and relevant (multiword) terms appearing in the input text. The system is evaluated in several scenarios, including the popular Reuters-21578 news corpus for comparison to other approaches, and categorization using IPTC metadata, EUROVOC thesaurus and others. Results show that this approach achieves a precision that is comparable to top ranked methods, with the added value that it does not require a demanding human expert workload to train

An evolutionary algorithm for the surface structure problem

Many macroscopic properties: hardness, corrosion, catalytic activity, etc. are directly related to the surface structure, that is, to the position and chemical identity of the outermost atoms of the material. Current experimental techniques for its determination produce a “signature” from which the structure must be inferred by solving an inverse problem: a solution is proposed, its corresponding signature computed and then compared to the experiment. This is a challenging optimization problem where the search space and the number of local minima grows exponentially with the number of atoms, hence its solution cannot be achieved for arbitrarily large structures. Nowadays, it is solved by using a mixture of human knowledge and local search techniques: an expert proposes a solution that is refined using a local minimizer. If the outcome does not fit the experiment, a new solution must be proposed again. Solving a small surface can take from days to weeks of this trial and error method. Here we describe our ongoing work in its solution. We use an hybrid algorithm that mixes evolutionary techniques with trusted region methods and reuses knowledge gained during the execution to avoid repeated search of structures. Its parallelization produces good results even when not requiring the gathering of the full population, hence it can be used in loosely coupled environments such as grids. With this algorithm, the solution of test cases that previously took weeks of expert time can be automatically solved in a day or two of uniprocessor time.

Hybrid incremental modeling based on least squares and fuzzy K-NN for monitoring tool wear in turning processes

There is now an emerging need for an efficient modeling strategy to develop a new generation of monitoring systems. One method of approaching the modeling of complex processes is to obtain a global model. It should be able to capture the basic or general behavior of the system, by means of a linear or quadratic regression, and then superimpose a local model on it that can capture the localized nonlinearities of the system. In this paper, a novel method based on a hybrid incremental modeling approach is designed and applied for tool wear detection in turning processes. It involves a two-step iterative process that combines a global model with a local model to take advantage of their underlying, complementary capacities. Thus, the first step constructs a global model using a least squares regression. A local model using the fuzzy k-nearest-neighbors smoothing algorithm is obtained in the second step. A comparative study then demonstrates that the hybrid incremental model provides better error-based performance indices for detecting tool wear than a transductive neurofuzzy model and an inductive neurofuzzy model.

Auto-localization algorithm for local positioning systems

This paper studies the problem of determining the position of beacon nodes in Local Positioning Systems (LPSs), for which there are no inter-beacon distance measurements available and neither the mobile node nor any of the stationary nodes have positioning or odometry information. The common solution is implemented using a mobile node capable of measuring its distance to the stationary beacon nodes within a sensing radius. Many authors have implemented heuristic methods based on optimization algorithms to solve the problem. However, such methods require a good initial estimation of the node positions in order to find the correct solution. In this paper we present a new method to calculate the inter-beacon distances, and hence the beacons positions, based in the linearization of the trilateration equations into a closed-form solution which does not require any approximate initial estimation. The simulations and field evaluations show a good estimation of the beacon node positions.

A guided heuristic approach to the MEG inverse problem.

There is general agreement within the scientific community in considering Biology as the science with more potential to develop in the XXI century. This is due to several reasons, but probably the most important one is the state of development of the rest of experimental and technological sciences. In this context, there are a very rich variety of mathematical tools, physical techniques and computer resources that permit to do biological experiments that were unbelievable only a few years ago. Biology is nowadays taking advantage of all these newly developed technologies, which are been applied to life sciences opening new research fields and helping to give new insights in many biological problems. Consequently, biologists have improved a lot their knowledge in many key areas as human function and human diseases. However there is one human organ that is still barely understood compared with the rest: The human brain. The understanding of the human brain is one of the main challenges of the XXI century. In this regard, it is considered a strategic research field for the European Union and the USA. Thus, there is a big interest in applying new experimental techniques for the study of brain function. Magnetoencephalography (MEG) is one of these novel techniques that are currently applied for mapping the brain activity1. This technique has important advantages compared to the metabolic-based brain imagining techniques like Functional Magneto Resonance Imaging2 (fMRI). The main advantage is that MEG has a higher time resolution than fMRI. Another benefit of MEG is that it is a patient friendly clinical technique. The measure is performed with a wireless set up and the patient is not exposed to any radiation. Although MEG is widely applied in clinical studies, there are still open issues regarding data analysis. The present work deals with the solution of the inverse problem in MEG, which is the most controversial and uncertain part of the analysis process3. This question is addressed using several variations of a new solving algorithm based in a heuristic method. The performance of those methods is analyzed by applying them to several test cases with known solutions and comparing those solutions with the ones provided by our methods.

A genetic algorithm for slope stability analyses with concave slip surfaces using custom operators

Heuristic methods are popular tools to find critical slip surfaces in slope stability analyses. A new genetic algorithm (GA) is proposed in this work that has a standard structure but a novel encoding and generation of individuals with custom-designed operators for mutation and crossover that produce kinematically feasible slip surfaces with a high probability. In addition, new indices to assess the efficiency of operators in their search for the minimum factor of safety (FS) are proposed. The proposed GA is applied to traditional benchmark examples from the literature, as well as to a new practical example. Results show that the proposed GA is reliable, flexible and robust: it provides good minimum FS estimates that are not very sensitive to the number of nodes and that are very similar for different replications