A proposal for improving safety in construction projects by strengthening coordinators’ competencies in health and safety issues

The figure of the coordinator in health and safety issues in the construction sector first appeared in our legislation through the incorporation of the European Directives (in our case Royal Decree 1627/97 on the minimum health and safety regulations in construction works), and is viewed differently in different countries of the European Union regarding the way they are hired and their role in the construction industry. Coordinating health and safety issues is also a management process that requires certain competencies that are not only based on technical or professional training, but which, taking account of the work environment, require the use of strategies and tools that are related to experience and personal skills. Through a piece of research that took account of expert opinions in the matter, we have found which competencies need to be possessed by the health and safety coordinator in order to improve the safety in the works they are coordinating. The conclusions of the analyses performed using the appropriate statistical methods (comparing means and multivariate analysis techniques), will enable training programmes to be designed and ensure that the health and safety coordinators selected have the competencies required to carry out their duties.

The role of residual stresses in the performance and durability of prestressing steel wires

Residual stresses developed during wire drawing influence the mechanical behavior and durability of steel wires used for prestressed concrete structures, particularly the shape of the stress–strain curve, stress relaxation losses, fatigue life, and environmental cracking susceptibility. The availability of general purpose finite element analysis tools and powerful diffraction techniques (X-rays and neutrons) has made it possible to predict and measure accurately residual stress fields in cold-drawn steel wires. Work carried out in this field in the past decade, shows the prospects and limitations of residual stress measurement, how the stress relaxation losses and environmentally-assisted cracking are correlated with the profile of residual stresses and how the performance of steel wires can be improved by modifying such a stress profile

Using clustering techniques for intelligent camera-based user interfaces

The area of Human-Machine Interface is growing fast due to its high importance in all technological systems. The basic idea behind designing human-machine interfaces is to enrich the communication with the technology in a natural and easy way. Gesture interfaces are a good example of transparent interfaces. Such interfaces must identify properly the action the user wants to perform, so the proper gesture recognition is of the highest importance. However, most of the systems based on gesture recognition use complex methods requiring high-resource devices. In this work, we propose to model gestures capturing their temporal properties, which significantly reduce storage requirements, and use clustering techniques, namely self-organizing maps and unsupervised genetic algorithm, for their classification. We further propose to train a certain number of algorithms with different parameters and combine their decision using majority voting in order to decrease the false positive rate. The main advantage of the approach is its simplicity, which enables the implementation using devices with limited resources, and therefore low cost. The testing results demonstrate its high potential.

Power management techniques in an FPGA-Based WSN node for high performance application

In this work, the power management techniques implemented in a high-performance node for Wireless Sensor Networks (WSN) based on a RAM-based FPGA are presented. This new node custom architecture is intended for high-end WSN applications that include complex sensor management like video cameras, high compute demanding tasks such as image encoding or robust encryption, and/or higher data bandwidth needs. In the case of these complex processing tasks, yet maintaining low power design requirements, it can be shown that the combination of different techniques such as extensive HW algorithm mapping, smart management of power islands to selectively switch on and off components, smart and low-energy partial reconfiguration, an adequate set of save energy modes and wake up options, all combined, may yield energy results that may compete and improve energy usage of typical low power microcontrollers used in many WSN node architectures. Actually, results show that higher complexity tasks are in favor of HW based platforms, while the flexibility achieved by dynamic and partial reconfiguration techniques could be comparable to SW based solutions.

Detection and Tracking of Dynamic Objects. A MultiRobot approach to Critical Infrastructures Surveillance.

Despite that Critical Infrastructures (CIs) security and surveillance are a growing concern for many countries and companies, Multi Robot Systems (MRSs) have not been yet broadly used in this type of facilities. This dissertation presents a novel study of the challenges arisen by the implementation of this type of systems and proposes solutions to specific problems. First, a comprehensive analysis of different types of CIs has been carried out, emphasizing the influence of the different characteristics of the facilities in the design of a security and surveillance MRS. One of the most important needs for the surveillance of a CI is the detection of intruders. From a technical point of view this problem can be abstracted as equivalent to the Detection and Tracking of Mobile Objects (DATMO). This dissertation proposes algorithms to solve this specific problem in a CI environment. Using 3D range images of the environment as input data, two detection algorithms for ground robots have been developed. These detection algorithms provide a list of moving objects in the robot detection area. Direct image differentiation and computer vision techniques are used when the robot is static. Alternatively, multi-layer ground reconstructions are compared to detect the dynamic objects when the robot is moving. Since CIs usually spread over large areas, it is very useful to incorporate aerial vehicles in the surveillance MRS. Therefore, a moving object detection algorithm for aerial vehicles has been also developed. This algorithm compares the real optical flow obtained from a down-face oriented camera with an artificial optical flow computed using a RANSAC based homography matrix. Two tracking algorithms have been developed to follow the moving objects trajectories. These algorithms can efficiently handle occlusions and crossings, as well as exchange information among robots. The multirobot tracking can be applied to any type of communication structure: centralized, decentralized or a combination of both. Even more, the developed tracking algorithms are independent of the detection algorithms and could be potentially used with other detection procedures or even with static sensors, such as cameras. In addition, using the 3D point clouds available to the robots, a relative localization algorithm has been developed to improve the position estimation of a given robot with observations from other robots. All the developed algorithms have been extensively tested in different simulated CIs using the Webots robotics simulator. Furthermore, the algorithms have also been validated with real robots operating in real scenarios. In conclusion, this dissertation presents a multirobot approach to Critical Infrastructure Surveillance, mainly focusing on Detecting and Tracking Dynamic Objects.

Multi-dimensional classification using Bayesian networks for stationary and evolving streaming data

Hoy en día, con la evolución continua y rápida de las tecnologías de la información y los dispositivos de computación, se recogen y almacenan continuamente grandes volúmenes de datos en distintos dominios y a través de diversas aplicaciones del mundo real. La extracción de conocimiento útil de una cantidad tan enorme de datos no se puede realizar habitualmente de forma manual, y requiere el uso de técnicas adecuadas de aprendizaje automático y de minería de datos. La clasificación es una de las técnicas más importantes que ha sido aplicada con éxito a varias áreas. En general, la clasificación se compone de dos pasos principales: en primer lugar, aprender un modelo de clasificación o clasificador a partir de un conjunto de datos de entrenamiento, y en segundo lugar, clasificar las nuevas instancias de datos utilizando el clasificador aprendido. La clasificación es supervisada cuando todas las etiquetas están presentes en los datos de entrenamiento (es decir, datos completamente etiquetados), semi-supervisada cuando sólo algunas etiquetas son conocidas (es decir, datos parcialmente etiquetados), y no supervisada cuando todas las etiquetas están ausentes en los datos de entrenamiento (es decir, datos no etiquetados). Además, aparte de esta taxonomía, el problema de clasificación se puede categorizar en unidimensional o multidimensional en función del número de variables clase, una o más, respectivamente; o también puede ser categorizado en estacionario o cambiante con el tiempo en función de las características de los datos y de la tasa de cambio subyacente. A lo largo de esta tesis, tratamos el problema de clasificación desde tres perspectivas diferentes, a saber, clasificación supervisada multidimensional estacionaria, clasificación semisupervisada unidimensional cambiante con el tiempo, y clasificación supervisada multidimensional cambiante con el tiempo. Para llevar a cabo esta tarea, hemos usado básicamente los clasificadores Bayesianos como modelos. La primera contribución, dirigiéndose al problema de clasificación supervisada multidimensional estacionaria, se compone de dos nuevos métodos de aprendizaje de clasificadores Bayesianos multidimensionales a partir de datos estacionarios. Los métodos se proponen desde dos puntos de vista diferentes. El primer método, denominado CB-MBC, se basa en una estrategia de envoltura de selección de variables que es voraz y hacia delante, mientras que el segundo, denominado MB-MBC, es una estrategia de filtrado de variables con una aproximación basada en restricciones y en el manto de Markov. Ambos métodos han sido aplicados a dos problemas reales importantes, a saber, la predicción de los inhibidores de la transcriptasa inversa y de la proteasa para el problema de infección por el virus de la inmunodeficiencia humana tipo 1 (HIV-1), y la predicción del European Quality of Life-5 Dimensions (EQ-5D) a partir de los cuestionarios de la enfermedad de Parkinson con 39 ítems (PDQ-39). El estudio experimental incluye comparaciones de CB-MBC y MB-MBC con los métodos del estado del arte de la clasificación multidimensional, así como con métodos comúnmente utilizados para resolver el problema de predicción de la enfermedad de Parkinson, a saber, la regresión logística multinomial, mínimos cuadrados ordinarios, y mínimas desviaciones absolutas censuradas. En ambas aplicaciones, los resultados han sido prometedores con respecto a la precisión de la clasificación, así como en relación al análisis de las estructuras gráficas que identifican interacciones conocidas y novedosas entre las variables. La segunda contribución, referida al problema de clasificación semi-supervisada unidimensional cambiante con el tiempo, consiste en un método nuevo (CPL-DS) para clasificar flujos de datos parcialmente etiquetados. Los flujos de datos difieren de los conjuntos de datos estacionarios en su proceso de generación muy rápido y en su aspecto de cambio de concepto. Es decir, los conceptos aprendidos y/o la distribución subyacente están probablemente cambiando y evolucionando en el tiempo, lo que hace que el modelo de clasificación actual sea obsoleto y deba ser actualizado. CPL-DS utiliza la divergencia de Kullback-Leibler y el método de bootstrapping para cuantificar y detectar tres tipos posibles de cambio: en las predictoras, en la a posteriori de la clase o en ambas. Después, si se detecta cualquier cambio, un nuevo modelo de clasificación se aprende usando el algoritmo EM; si no, el modelo de clasificación actual se mantiene sin modificaciones. CPL-DS es general, ya que puede ser aplicado a varios modelos de clasificación. Usando dos modelos diferentes, el clasificador naive Bayes y la regresión logística, CPL-DS se ha probado con flujos de datos sintéticos y también se ha aplicado al problema real de la detección de código malware, en el cual los nuevos ficheros recibidos deben ser continuamente clasificados en malware o goodware. Los resultados experimentales muestran que nuestro método es efectivo para la detección de diferentes tipos de cambio a partir de los flujos de datos parcialmente etiquetados y también tiene una buena precisión de la clasificación. Finalmente, la tercera contribución, sobre el problema de clasificación supervisada multidimensional cambiante con el tiempo, consiste en dos métodos adaptativos, a saber, Locally Adpative-MB-MBC (LA-MB-MBC) y Globally Adpative-MB-MBC (GA-MB-MBC). Ambos métodos monitorizan el cambio de concepto a lo largo del tiempo utilizando la log-verosimilitud media como métrica y el test de Page-Hinkley. Luego, si se detecta un cambio de concepto, LA-MB-MBC adapta el actual clasificador Bayesiano multidimensional localmente alrededor de cada nodo cambiado, mientras que GA-MB-MBC aprende un nuevo clasificador Bayesiano multidimensional. El estudio experimental realizado usando flujos de datos sintéticos multidimensionales indica los méritos de los métodos adaptativos propuestos. ABSTRACT Nowadays, with the ongoing and rapid evolution of information technology and computing devices, large volumes of data are continuously collected and stored in different domains and through various real-world applications. Extracting useful knowledge from such a huge amount of data usually cannot be performed manually, and requires the use of adequate machine learning and data mining techniques. Classification is one of the most important techniques that has been successfully applied to several areas. Roughly speaking, classification consists of two main steps: first, learn a classification model or classifier from an available training data, and secondly, classify the new incoming unseen data instances using the learned classifier. Classification is supervised when the whole class values are present in the training data (i.e., fully labeled data), semi-supervised when only some class values are known (i.e., partially labeled data), and unsupervised when the whole class values are missing in the training data (i.e., unlabeled data). In addition, besides this taxonomy, the classification problem can be categorized into uni-dimensional or multi-dimensional depending on the number of class variables, one or more, respectively; or can be also categorized into stationary or streaming depending on the characteristics of the data and the rate of change underlying it. Through this thesis, we deal with the classification problem under three different settings, namely, supervised multi-dimensional stationary classification, semi-supervised unidimensional streaming classification, and supervised multi-dimensional streaming classification. To accomplish this task, we basically used Bayesian network classifiers as models. The first contribution, addressing the supervised multi-dimensional stationary classification problem, consists of two new methods for learning multi-dimensional Bayesian network classifiers from stationary data. They are proposed from two different points of view. The first method, named CB-MBC, is based on a wrapper greedy forward selection approach, while the second one, named MB-MBC, is a filter constraint-based approach based on Markov blankets. Both methods are applied to two important real-world problems, namely, the prediction of the human immunodeficiency virus type 1 (HIV-1) reverse transcriptase and protease inhibitors, and the prediction of the European Quality of Life-5 Dimensions (EQ-5D) from 39-item Parkinson’s Disease Questionnaire (PDQ-39). The experimental study includes comparisons of CB-MBC and MB-MBC against state-of-the-art multi-dimensional classification methods, as well as against commonly used methods for solving the Parkinson’s disease prediction problem, namely, multinomial logistic regression, ordinary least squares, and censored least absolute deviations. For both considered case studies, results are promising in terms of classification accuracy as well as regarding the analysis of the learned MBC graphical structures identifying known and novel interactions among variables. The second contribution, addressing the semi-supervised uni-dimensional streaming classification problem, consists of a novel method (CPL-DS) for classifying partially labeled data streams. Data streams differ from the stationary data sets by their highly rapid generation process and their concept-drifting aspect. That is, the learned concepts and/or the underlying distribution are likely changing and evolving over time, which makes the current classification model out-of-date requiring to be updated. CPL-DS uses the Kullback-Leibler divergence and bootstrapping method to quantify and detect three possible kinds of drift: feature, conditional or dual. Then, if any occurs, a new classification model is learned using the expectation-maximization algorithm; otherwise, the current classification model is kept unchanged. CPL-DS is general as it can be applied to several classification models. Using two different models, namely, naive Bayes classifier and logistic regression, CPL-DS is tested with synthetic data streams and applied to the real-world problem of malware detection, where the new received files should be continuously classified into malware or goodware. Experimental results show that our approach is effective for detecting different kinds of drift from partially labeled data streams, as well as having a good classification performance. Finally, the third contribution, addressing the supervised multi-dimensional streaming classification problem, consists of two adaptive methods, namely, Locally Adaptive-MB-MBC (LA-MB-MBC) and Globally Adaptive-MB-MBC (GA-MB-MBC). Both methods monitor the concept drift over time using the average log-likelihood score and the Page-Hinkley test. Then, if a drift is detected, LA-MB-MBC adapts the current multi-dimensional Bayesian network classifier locally around each changed node, whereas GA-MB-MBC learns a new multi-dimensional Bayesian network classifier from scratch. Experimental study carried out using synthetic multi-dimensional data streams shows the merits of both proposed adaptive methods.

Localization and Tracking in Energy-Constrained Networks with Dynamic Connectivity

Systems used for target localization, such as goods, individuals, or animals, commonly rely on operational means to meet the final application demands. However, what would happen if some means were powered up randomly by harvesting systems? And what if those devices not randomly powered had their duty cycles restricted? Under what conditions would such an operation be tolerable in localization services? What if the references provided by nodes in a tracking problem were distorted? Moreover, there is an underlying topic common to the previous questions regarding the transfer of conceptual models to reality in field tests: what challenges are faced upon deploying a localization network that integrates energy harvesting modules? The application scenario of the system studied is a traditional herding environment of semi domesticated reindeer (Rangifer tarandus tarandus) in northern Scandinavia. In these conditions, information on approximate locations of reindeer is as important as environmental preservation. Herders also need cost-effective devices capable of operating unattended in, sometimes, extreme weather conditions. The analyses developed are worthy not only for the specific application environment presented, but also because they may serve as an approach to performance of navigation systems in absence of reasonably accurate references like the ones of the Global Positioning System (GPS). A number of energy-harvesting solutions, like thermal and radio-frequency harvesting, do not commonly provide power beyond one milliwatt. When they do, battery buffers may be needed (as it happens with solar energy) which may raise costs and make systems more dependent on environmental temperatures. In general, given our problem, a harvesting system is needed that be capable of providing energy bursts of, at least, some milliwatts. Many works on localization problems assume that devices have certain capabilities to determine unknown locations based on range-based techniques or fingerprinting which cannot be assumed in the approach considered herein. The system presented is akin to range-free techniques, but goes to the extent of considering very low node densities: most range-free techniques are, therefore, not applicable. Animal localization, in particular, uses to be supported by accurate devices such as GPS collars which deplete batteries in, maximum, a few days. Such short-life solutions are not particularly desirable in the framework considered. In tracking, the challenge may times addressed aims at attaining high precision levels from complex reliable hardware and thorough processing techniques. One of the challenges in this Thesis is the use of equipment with just part of its facilities in permanent operation, which may yield high input noise levels in the form of distorted reference points. The solution presented integrates a kinetic harvesting module in some nodes which are expected to be a majority in the network. These modules are capable of providing power bursts of some milliwatts which suffice to meet node energy demands. The usage of harvesting modules in the aforementioned conditions makes the system less dependent on environmental temperatures as no batteries are used in nodes with harvesters--it may be also an advantage in economic terms. There is a second kind of nodes. They are battery powered (without kinetic energy harvesters), and are, therefore, dependent on temperature and battery replacements. In addition, their operation is constrained by duty cycles in order to extend node lifetime and, consequently, their autonomy. There is, in turn, a third type of nodes (hotspots) which can be static or mobile. They are also battery-powered, and are used to retrieve information from the network so that it is presented to users. The system operational chain starts at the kinetic-powered nodes broadcasting their own identifier. If an identifier is received at a battery-powered node, the latter stores it for its records. Later, as the recording node meets a hotspot, its full record of detections is transferred to the hotspot. Every detection registry comprises, at least, a node identifier and the position read from its GPS module by the battery-operated node previously to detection. The characteristics of the system presented make the aforementioned operation own certain particularities which are also studied. First, identifier transmissions are random as they depend on movements at kinetic modules--reindeer movements in our application. Not every movement suffices since it must overcome a certain energy threshold. Second, identifier transmissions may not be heard unless there is a battery-powered node in the surroundings. Third, battery-powered nodes do not poll continuously their GPS module, hence localization errors rise even more. Let's recall at this point that such behavior is tight to the aforementioned power saving policies to extend node lifetime. Last, some time is elapsed between the instant an identifier random transmission is detected and the moment the user is aware of such a detection: it takes some time to find a hotspot. Tracking is posed as a problem of a single kinetically-powered target and a population of battery-operated nodes with higher densities than before in localization. Since the latter provide their approximate positions as reference locations, the study is again focused on assessing the impact of such distorted references on performance. Unlike in localization, distance-estimation capabilities based on signal parameters are assumed in this problem. Three variants of the Kalman filter family are applied in this context: the regular Kalman filter, the alpha-beta filter, and the unscented Kalman filter. The study enclosed hereafter comprises both field tests and simulations. Field tests were used mainly to assess the challenges related to power supply and operation in extreme conditions as well as to model nodes and some aspects of their operation in the application scenario. These models are the basics of the simulations developed later. The overall system performance is analyzed according to three metrics: number of detections per kinetic node, accuracy, and latency. The links between these metrics and the operational conditions are also discussed and characterized statistically. Subsequently, such statistical characterization is used to forecast performance figures given specific operational parameters. In tracking, also studied via simulations, nonlinear relationships are found between accuracy and duty cycles and cluster sizes of battery-operated nodes. The solution presented may be more complex in terms of network structure than existing solutions based on GPS collars. However, its main gain lies on taking advantage of users' error tolerance to reduce costs and become more environmentally friendly by diminishing the potential amount of batteries that can be lost. Whether it is applicable or not depends ultimately on the conditions and requirements imposed by users' needs and operational environments, which is, as it has been explained, one of the topics of this Thesis.

Using array based techniques for noise source identification in cars

Una de las principales causas del ruido en nuestras ciudades es el tráfico rodado. El ruido generado por los vehículos no es sólo debido al motor, sino que existen diversas fuentes de ruido en los mismos, entre las que se puede destacar el ruido de rodadura. Para localizar las causas del ruido e identificar las principales fuentes del mismo se han utilizado en diversos estudios las técnicas de coherencia y las técnicas basadas en arrays. Sin embargo, en la bibliografía existente, no es habitual encontrar el uso de estas técnicas en el sector automovilístico. En esta tesis se parte de la premisa de la posibilidad de usar estas técnicas de medida en coches, para demostrar a la largo de la misma su factibilidad y su bondad para evaluar las fuentes de ruido en dos condiciones distintas: cuando el coche está parado y cuando está en movimiento. Como técnica de coherencia se elige la de Intensidad Selectiva, utilizándose la misma para evaluar la coherencia existente entre el ruido que llega a los oídos del conductor y la intensidad radiada por distintos puntos del motor. Para la localización de fuentes de ruido, las técnicas basadas en array son las que mejores resultados ofrecen. Statistically Optimized Near-field Acoustical Holography (SONAH) es la técnica elegida para la localización y caracterización de las fuentes de ruido en el motor a baja frecuencia. En cambio, Beamforming es la técnica seleccionada para el caso de media-alta frecuencia y para la evaluación de las fuentes de ruido cuando el coche se encuentra en movimiento. Las técnicas propuestas no sólo pueden utilizarse en medidas reales, sino que además proporcionan abundante información y frecen una gran versatilidad a la hora de caracterizar fuentes de ruido. ABSTRACT One of the most important noise causes in our cities is the traffic. The noise generated by the vehicles is not only due to the engine, but there are some other noise sources. Among them the tyre/road noise can be highlighted. Coherence and array based techniques have been used in some research to locate the noise causes and identify the main noise sources. Nevertheless, it is not usual in the literature to find the application of this kind of techniques in the car sector. This Thesis starts taking into account the possibility of using this kind of measurement techniques in cars, to demonstrate their feasability and their quality to evaluate the noise sources under two different conditions: when the car is stopped and when it is in movement. Selective Intensity was chosen as coherence technique, evaluating the coherence between the noise in the driver’s ears and the intensity radiated in different points of the engine. Array based techniques carry out the best results to noise source location. Statistically Optimized Near-field Acoustical Holography (SONAH) is the measurement technique chosen for noise source location and characterization in the engine at low frequency. On the other hand, Beamforming is the technique chosen in the case of medium-high frequency and to characterize the noise sources when the car is in movement. The proposed techniques not only can be used in actual measurements, but also provide a lot of information and are very versatile to noise source characterization.

Multidimensional, multislaces, and multiscale MRI for quality evaluation in food.

InsideFood explicitly aims at measuring food microstructure, the spatial distribution of food components within foods, with state of the art tomographic, spectroscopic and texture measurement techniques including X-ray micro-and nano CT, MRI,OCT, NMR, TRS and SRS, and acoustic emission. Nutritional quality (sugar and gluten free cereal products), sensory quality (texture of all foods) and safety (foreign material detection in cereal products) are considered. Online and inline techniques including NMR, MRI, TRS, SRS and X-ray imaging to visualise and monitor structure will be developed.

Digital Coherent Receivers and Advanced Optical Modulation Formats in 100 and 200 Gb/s DWDM Systems

Recent advances in coherent optical receivers is reviewed. Digital-Signal-Processing (DSP) based phase and polarization management techniques make coherent detection robust and feasible. With coherent detection, the complex field of the received optical signal is fully recovered, allowing compensation of linear and nonlinear optical impairments including chromatic dispersion (CD) and polarization-mode dispersion (PMD) using digital filters. Coherent detection and advanced optical modulation formats have become a key ingredient to the design of modern dense wavelength-division multiplexed (DWDM) optical broadband networks. In this paper, firstly we present the different subsystems of a digital coherent optical receiver, and secondly, we will compare the performance of some multi-level and multi-dimensional modulation formats in some physical impairments and in high spectral-efficiency (SE) and high-capacity DWDM transmissions, simulating the DSP with Matlab and the optical network performance with OptiSystem software.

Nanoinformatics approaches for information extraction and text mining in nanomedical research texts

La nanotecnología es un área de investigación de reciente creación que trata con la manipulación y el control de la materia con dimensiones comprendidas entre 1 y 100 nanómetros. A escala nanométrica, los materiales exhiben fenómenos físicos, químicos y biológicos singulares, muy distintos a los que manifiestan a escala convencional. En medicina, los compuestos miniaturizados a nanoescala y los materiales nanoestructurados ofrecen una mayor eficacia con respecto a las formulaciones químicas tradicionales, así como una mejora en la focalización del medicamento hacia la diana terapéutica, revelando así nuevas propiedades diagnósticas y terapéuticas. A su vez, la complejidad de la información a nivel nano es mucho mayor que en los niveles biológicos convencionales (desde el nivel de población hasta el nivel de célula) y, por tanto, cualquier flujo de trabajo en nanomedicina requiere, de forma inherente, estrategias de gestión de información avanzadas. Desafortunadamente, la informática biomédica todavía no ha proporcionado el marco de trabajo que permita lidiar con estos retos de la información a nivel nano, ni ha adaptado sus métodos y herramientas a este nuevo campo de investigación. En este contexto, la nueva área de la nanoinformática pretende detectar y establecer los vínculos existentes entre la medicina, la nanotecnología y la informática, fomentando así la aplicación de métodos computacionales para resolver las cuestiones y problemas que surgen con la información en la amplia intersección entre la biomedicina y la nanotecnología. Las observaciones expuestas previamente determinan el contexto de esta tesis doctoral, la cual se centra en analizar el dominio de la nanomedicina en profundidad, así como en el desarrollo de estrategias y herramientas para establecer correspondencias entre las distintas disciplinas, fuentes de datos, recursos computacionales y técnicas orientadas a la extracción de información y la minería de textos, con el objetivo final de hacer uso de los datos nanomédicos disponibles. El autor analiza, a través de casos reales, alguna de las tareas de investigación en nanomedicina que requieren o que pueden beneficiarse del uso de métodos y herramientas nanoinformáticas, ilustrando de esta forma los inconvenientes y limitaciones actuales de los enfoques de informática biomédica a la hora de tratar con datos pertenecientes al dominio nanomédico. Se discuten tres escenarios diferentes como ejemplos de actividades que los investigadores realizan mientras llevan a cabo su investigación, comparando los contextos biomédico y nanomédico: i) búsqueda en la Web de fuentes de datos y recursos computacionales que den soporte a su investigación; ii) búsqueda en la literatura científica de resultados experimentales y publicaciones relacionadas con su investigación; iii) búsqueda en registros de ensayos clínicos de resultados clínicos relacionados con su investigación. El desarrollo de estas actividades requiere el uso de herramientas y servicios informáticos, como exploradores Web, bases de datos de referencias bibliográficas indexando la literatura biomédica y registros online de ensayos clínicos, respectivamente. Para cada escenario, este documento proporciona un análisis detallado de los posibles obstáculos que pueden dificultar el desarrollo y el resultado de las diferentes tareas de investigación en cada uno de los dos campos citados (biomedicina y nanomedicina), poniendo especial énfasis en los retos existentes en la investigación nanomédica, campo en el que se han detectado las mayores dificultades. El autor ilustra cómo la aplicación de metodologías provenientes de la informática biomédica a estos escenarios resulta efectiva en el dominio biomédico, mientras que dichas metodologías presentan serias limitaciones cuando son aplicadas al contexto nanomédico. Para abordar dichas limitaciones, el autor propone un enfoque nanoinformático, original, diseñado específicamente para tratar con las características especiales que la información presenta a nivel nano. El enfoque consiste en un análisis en profundidad de la literatura científica y de los registros de ensayos clínicos disponibles para extraer información relevante sobre experimentos y resultados en nanomedicina —patrones textuales, vocabulario en común, descriptores de experimentos, parámetros de caracterización, etc.—, seguido del desarrollo de mecanismos para estructurar y analizar dicha información automáticamente. Este análisis concluye con la generación de un modelo de datos de referencia (gold standard) —un conjunto de datos de entrenamiento y de test anotados manualmente—, el cual ha sido aplicado a la clasificación de registros de ensayos clínicos, permitiendo distinguir automáticamente los estudios centrados en nanodrogas y nanodispositivos de aquellos enfocados a testear productos farmacéuticos tradicionales. El presente trabajo pretende proporcionar los métodos necesarios para organizar, depurar, filtrar y validar parte de los datos nanomédicos existentes en la actualidad a una escala adecuada para la toma de decisiones. Análisis similares para otras tareas de investigación en nanomedicina ayudarían a detectar qué recursos nanoinformáticos se requieren para cumplir los objetivos actuales en el área, así como a generar conjunto de datos de referencia, estructurados y densos en información, a partir de literatura y otros fuentes no estructuradas para poder aplicar nuevos algoritmos e inferir nueva información de valor para la investigación en nanomedicina. ABSTRACT Nanotechnology is a research area of recent development that deals with the manipulation and control of matter with dimensions ranging from 1 to 100 nanometers. At the nanoscale, materials exhibit singular physical, chemical and biological phenomena, very different from those manifested at the conventional scale. In medicine, nanosized compounds and nanostructured materials offer improved drug targeting and efficacy with respect to traditional formulations, and reveal novel diagnostic and therapeutic properties. Nevertheless, the complexity of information at the nano level is much higher than the complexity at the conventional biological levels (from populations to the cell). Thus, any nanomedical research workflow inherently demands advanced information management. Unfortunately, Biomedical Informatics (BMI) has not yet provided the necessary framework to deal with such information challenges, nor adapted its methods and tools to the new research field. In this context, the novel area of nanoinformatics aims to build new bridges between medicine, nanotechnology and informatics, allowing the application of computational methods to solve informational issues at the wide intersection between biomedicine and nanotechnology. The above observations determine the context of this doctoral dissertation, which is focused on analyzing the nanomedical domain in-depth, and developing nanoinformatics strategies and tools to map across disciplines, data sources, computational resources, and information extraction and text mining techniques, for leveraging available nanomedical data. The author analyzes, through real-life case studies, some research tasks in nanomedicine that would require or could benefit from the use of nanoinformatics methods and tools, illustrating present drawbacks and limitations of BMI approaches to deal with data belonging to the nanomedical domain. Three different scenarios, comparing both the biomedical and nanomedical contexts, are discussed as examples of activities that researchers would perform while conducting their research: i) searching over the Web for data sources and computational resources supporting their research; ii) searching the literature for experimental results and publications related to their research, and iii) searching clinical trial registries for clinical results related to their research. The development of these activities will depend on the use of informatics tools and services, such as web browsers, databases of citations and abstracts indexing the biomedical literature, and web-based clinical trial registries, respectively. For each scenario, this document provides a detailed analysis of the potential information barriers that could hamper the successful development of the different research tasks in both fields (biomedicine and nanomedicine), emphasizing the existing challenges for nanomedical research —where the major barriers have been found. The author illustrates how the application of BMI methodologies to these scenarios can be proven successful in the biomedical domain, whilst these methodologies present severe limitations when applied to the nanomedical context. To address such limitations, the author proposes an original nanoinformatics approach specifically designed to deal with the special characteristics of information at the nano level. This approach consists of an in-depth analysis of the scientific literature and available clinical trial registries to extract relevant information about experiments and results in nanomedicine —textual patterns, common vocabulary, experiment descriptors, characterization parameters, etc.—, followed by the development of mechanisms to automatically structure and analyze this information. This analysis resulted in the generation of a gold standard —a manually annotated training or reference set—, which was applied to the automatic classification of clinical trial summaries, distinguishing studies focused on nanodrugs and nanodevices from those aimed at testing traditional pharmaceuticals. The present work aims to provide the necessary methods for organizing, curating and validating existing nanomedical data on a scale suitable for decision-making. Similar analysis for different nanomedical research tasks would help to detect which nanoinformatics resources are required to meet current goals in the field, as well as to generate densely populated and machine-interpretable reference datasets from the literature and other unstructured sources for further testing novel algorithms and inferring new valuable information for nanomedicine.

Automatic physical activity pattern recognition and meal intake modelling in type 1 diabetes for their inclusion in artificial pancreas systems

La diabetes comprende un conjunto de enfermedades metabólicas que se caracterizan por concentraciones de glucosa en sangre anormalmente altas. En el caso de la diabetes tipo 1 (T1D, por sus siglas en inglés), esta situación es debida a una ausencia total de secreción endógena de insulina, lo que impide a la mayoría de tejidos usar la glucosa. En tales circunstancias, se hace necesario el suministro exógeno de insulina para preservar la vida del paciente; no obstante, siempre con la precaución de evitar caídas agudas de la glucemia por debajo de los niveles recomendados de seguridad. Además de la administración de insulina, las ingestas y la actividad física son factores fundamentales que influyen en la homeostasis de la glucosa. En consecuencia, una gestión apropiada de la T1D debería incorporar estos dos fenómenos fisiológicos, en base a una identificación y un modelado apropiado de los mismos y de sus sorrespondientes efectos en el balance glucosa-insulina. En particular, los sistemas de páncreas artificial –ideados para llevar a cabo un control automático de los niveles de glucemia del paciente– podrían beneficiarse de la integración de esta clase de información. La primera parte de esta tesis doctoral cubre la caracterización del efecto agudo de la actividad física en los perfiles de glucosa. Con este objetivo se ha llevado a cabo una revisión sistemática de la literatura y meta-análisis que determinen las respuestas ante varias modalidades de ejercicio para pacientes con T1D, abordando esta caracterización mediante unas magnitudes que cuantifican las tasas de cambio en la glucemia a lo largo del tiempo. Por otro lado, una identificación fiable de los periodos con actividad física es un requisito imprescindible para poder proveer de esa información a los sistemas de páncreas artificial en condiciones libres y ambulatorias. Por esta razón, la segunda parte de esta tesis está enfocada a la propuesta y evaluación de un sistema automático diseñado para reconocer periodos de actividad física, clasificando su nivel de intensidad (ligera, moderada o vigorosa); así como, en el caso de periodos vigorosos, identificando también la modalidad de ejercicio (aeróbica, mixta o de fuerza). En este sentido, ambos aspectos tienen una influencia específica en el mecanismo metabólico que suministra la energía para llevar a cabo el ejercicio y, por tanto, en las respuestas glucémicas en T1D. En este trabajo se aplican varias combinaciones de técnicas de aprendizaje máquina y reconocimiento de patrones sobre la fusión multimodal de señales de acelerometría y ritmo cardíaco, las cuales describen tanto aspectos mecánicos del movimiento como la respuesta fisiológica del sistema cardiovascular ante el ejercicio. Después del reconocimiento de patrones se incorpora también un módulo de filtrado temporal para sacar partido a la considerable coherencia temporal presente en los datos, una redundancia que se origina en el hecho de que en la práctica, las tendencias en cuanto a actividad física suelen mantenerse estables a lo largo de cierto tiempo, sin fluctuaciones rápidas y repetitivas. El tercer bloque de esta tesis doctoral aborda el tema de las ingestas en el ámbito de la T1D. En concreto, se propone una serie de modelos compartimentales y se evalúan éstos en función de su capacidad para describir matemáticamente el efecto remoto de las concetraciones plasmáticas de insulina exógena sobre las tasas de eleiminación de la glucosa atribuible a la ingesta; un aspecto hasta ahora no incorporado en los principales modelos de paciente para T1D existentes en la literatura. Los datos aquí utilizados se obtuvieron gracias a un experimento realizado por el Institute of Metabolic Science (Universidad de Cambridge, Reino Unido) con 16 pacientes jóvenes. En el experimento, de tipo ‘clamp’ con objetivo variable, se replicaron los perfiles individuales de glucosa, según lo observado durante una visita preliminar tras la ingesta de una cena con o bien alta carga glucémica, o bien baja. Los seis modelos mecanísticos evaluados constaban de: a) submodelos de doble compartimento para las masas de trazadores de glucosa, b) un submodelo de único compartimento para reflejar el efecto remoto de la insulina, c) dos tipos de activación de este mismo efecto remoto (bien lineal, bien con un punto de corte), y d) diversas condiciones iniciales. ABSTRACT Diabetes encompasses a series of metabolic diseases characterized by abnormally high blood glucose concentrations. In the case of type 1 diabetes (T1D), this situation is caused by a total absence of endogenous insulin secretion, which impedes the use of glucose by most tissues. In these circumstances, exogenous insulin supplies are necessary to maintain patient’s life; although caution is always needed to avoid acute decays in glycaemia below safe levels. In addition to insulin administrations, meal intakes and physical activity are fundamental factors influencing glucose homoeostasis. Consequently, a successful management of T1D should incorporate these two physiological phenomena, based on an appropriate identification and modelling of these events and their corresponding effect on the glucose-insulin balance. In particular, artificial pancreas systems –designed to perform an automated control of patient’s glycaemia levels– may benefit from the integration of this type of information. The first part of this PhD thesis covers the characterization of the acute effect of physical activity on glucose profiles. With this aim, a systematic review of literature and metaanalyses are conduced to determine responses to various exercise modalities in patients with T1D, assessed via rates-of-change magnitudes to quantify temporal variations in glycaemia. On the other hand, a reliable identification of physical activity periods is an essential prerequisite to feed artificial pancreas systems with information concerning exercise in ambulatory, free-living conditions. For this reason, the second part of this thesis focuses on the proposal and evaluation of an automatic system devised to recognize physical activity, classifying its intensity level (light, moderate or vigorous) and for vigorous periods, identifying also its exercise modality (aerobic, mixed or resistance); since both aspects have a distinctive influence on the predominant metabolic pathway involved in fuelling exercise, and therefore, in the glycaemic responses in T1D. Various combinations of machine learning and pattern recognition techniques are applied on the fusion of multi-modal signal sources, namely: accelerometry and heart rate measurements, which describe both mechanical aspects of movement and the physiological response of the cardiovascular system to exercise. An additional temporal filtering module is incorporated after recognition in order to exploit the considerable temporal coherence (i.e. redundancy) present in data, which stems from the fact that in practice, physical activity trends are often maintained stable along time, instead of fluctuating rapid and repeatedly. The third block of this PhD thesis addresses meal intakes in the context of T1D. In particular, a number of compartmental models are proposed and compared in terms of their ability to describe mathematically the remote effect of exogenous plasma insulin concentrations on the disposal rates of meal-attributable glucose, an aspect which had not yet been incorporated to the prevailing T1D patient models in literature. Data were acquired in an experiment conduced at the Institute of Metabolic Science (University of Cambridge, UK) on 16 young patients. A variable-target glucose clamp replicated their individual glucose profiles, observed during a preliminary visit after ingesting either a high glycaemic-load or a low glycaemic-load evening meal. The six mechanistic models under evaluation here comprised: a) two-compartmental submodels for glucose tracer masses, b) a single-compartmental submodel for insulin’s remote effect, c) two types of activations for this remote effect (either linear or with a ‘cut-off’ point), and d) diverse forms of initial conditions.

Characterising social media users by gender and place of residence

Characterising users through demographic attributes is a necessary step before conducting opinion surveys from information published by such users in social media. In this paper, we describe, compare and evaluate different techniques for the identification of the attributes "gender"' and "place of residence" by mining the metadata associated to the users, the content published and shared by themselves, and their friendship networks. The results obtained show that the social network is a valuable source of information for obtaining the sociodemographic attributes of single users.

A multi-level analysis of innovation diffusion: The case of solar photovoltaic systems in Germany and the disillusion of grid parity

¿Suministrarán las fuentes de energía renovables toda la energía que el mundo necesita algún día? Algunos argumentan que sí, mientras que otros dicen que no. Sin embargo, en algunas regiones del mundo, la producción de electricidad a través de fuentes de energía renovables ya está en una etapa prometedora de desarrollo en la que su costo de generación de electricidad compite con fuentes de electricidad convencionales, como por ejemplo la paridad de red. Este logro ha sido respaldado por el aumento de la eficiencia de la tecnología, la reducción de los costos de producción y, sobre todo, los años de intervenciones políticas de apoyo financiero. La difusión de los sistemas solares fotovoltaicos (PV) en Alemania es un ejemplo relevante. Alemania no sólo es el país líder en términos de capacidad instalada de sistemas fotovoltaicos (PV) en todo el mundo, sino también uno de los países pioneros donde la paridad de red se ha logrado recientemente. No obstante, podría haber una nube en el horizonte. La tasa de difusión ha comenzado a declinar en muchas regiones. Además, las empresas solares locales – que se sabe son importantes impulsores de la difusión – han comenzado a enfrentar dificultades para manejar sus negocios. Estos acontecimientos plantean algunas preguntas importantes: ¿Es ésta una disminución temporal en la difusión? ¿Los adoptantes continuarán instalando sistemas fotovoltaicos? ¿Qué pasa con los modelos de negocio de las empresas solares locales? Con base en el caso de los sistemas fotovoltaicos en Alemania a través de un análisis multinivel y dos revisiones literarias complementarias, esta tesis doctoral extiende el debate proporcionando riqueza múltiple de datos empíricos en un conocimiento de contexto limitado. El primer análisis se basa en la perspectiva del adoptante, que explora el nivel "micro" y el proceso social que subyace a la adopción de los sistemas fotovoltaicos. El segundo análisis es una perspectiva a nivel de empresa, que explora los modelos de negocio de las empresas y sus roles impulsores en la difusión de los sistemas fotovoltaicos. El tercero análisis es una perspectiva regional, la cual explora el nivel "meso", el proceso social que subyace a la adopción de sistemas fotovoltaicos y sus técnicas de modelado. Los resultados incluyen implicaciones tanto para académicos como políticos, no sólo sobre las innovaciones en energía renovable relativas a la paridad de red, sino también, de manera inductiva, sobre las innovaciones ambientales impulsadas por las políticas que logren la competitividad de costes. ABSTRACT Will renewable energy sources supply all of the world energy needs one day? Some argue yes, while others say no. However, in some regions of the world, the electricity production through renewable energy sources is already at a promising stage of development at which their electricity generation costs compete with conventional electricity sources’, i.e., grid parity. This achievement has been underpinned by the increase of technology efficiency, reduction of production costs and, above all, years of policy interventions of providing financial support. The diffusion of solar photovoltaic (PV) systems in Germany is an important frontrunner case in point. Germany is not only the top country in terms of installed PV systems’ capacity worldwide but also one of the pioneer countries where the grid parity has recently been achieved. However, there might be a cloud on the horizon. The diffusion rate has started to decline in many regions. In addition, local solar firms – which are known to be important drivers of diffusion – have started to face difficulties to run their businesses. These developments raise some important questions: Is this a temporary decline on diffusion? Will adopters continue to install PV systems? What about the business models of the local solar firms? Based on the case of PV systems in Germany through a multi-level analysis and two complementary literature reviews, this PhD Dissertation extends the debate by providing multiple wealth of empirical details in a context-limited knowledge. The first analysis is based on the adopter perspective, which explores the “micro” level and the social process underlying the adoption of PV systems. The second one is a firm-level perspective, which explores the business models of firms and their driving roles in diffusion of PV systems. The third one is a regional perspective, which explores the “meso” level, i.e., the social process underlying the adoption of PV systems and its modeling techniques. The results include implications for both scholars and policymakers, not only about renewable energy innovations at grid parity, but also in an inductive manner, about policy-driven environmental innovations that achieve the cost competiveness.

Contributions to Speech Analytics based on Speech Recognition and Topic Identification

La última década ha sido testigo de importantes avances en el campo de la tecnología de reconocimiento de voz. Los sistemas comerciales existentes actualmente poseen la capacidad de reconocer habla continua de múltiples locutores, consiguiendo valores aceptables de error, y sin la necesidad de realizar procedimientos explícitos de adaptación. A pesar del buen momento que vive esta tecnología, el reconocimiento de voz dista de ser un problema resuelto. La mayoría de estos sistemas de reconocimiento se ajustan a dominios particulares y su eficacia depende de manera significativa, entre otros muchos aspectos, de la similitud que exista entre el modelo de lenguaje utilizado y la tarea específica para la cual se está empleando. Esta dependencia cobra aún más importancia en aquellos escenarios en los cuales las propiedades estadísticas del lenguaje varían a lo largo del tiempo, como por ejemplo, en dominios de aplicación que involucren habla espontánea y múltiples temáticas. En los últimos años se ha evidenciado un constante esfuerzo por mejorar los sistemas de reconocimiento para tales dominios. Esto se ha hecho, entre otros muchos enfoques, a través de técnicas automáticas de adaptación. Estas técnicas son aplicadas a sistemas ya existentes, dado que exportar el sistema a una nueva tarea o dominio puede requerir tiempo a la vez que resultar costoso. Las técnicas de adaptación requieren fuentes adicionales de información, y en este sentido, el lenguaje hablado puede aportar algunas de ellas. El habla no sólo transmite un mensaje, también transmite información acerca del contexto en el cual se desarrolla la comunicación hablada (e.g. acerca del tema sobre el cual se está hablando). Por tanto, cuando nos comunicamos a través del habla, es posible identificar los elementos del lenguaje que caracterizan el contexto, y al mismo tiempo, rastrear los cambios que ocurren en estos elementos a lo largo del tiempo. Esta información podría ser capturada y aprovechada por medio de técnicas de recuperación de información (information retrieval) y de aprendizaje de máquina (machine learning). Esto podría permitirnos, dentro del desarrollo de mejores sistemas automáticos de reconocimiento de voz, mejorar la adaptación de modelos del lenguaje a las condiciones del contexto, y por tanto, robustecer al sistema de reconocimiento en dominios con condiciones variables (tales como variaciones potenciales en el vocabulario, el estilo y la temática). En este sentido, la principal contribución de esta Tesis es la propuesta y evaluación de un marco de contextualización motivado por el análisis temático y basado en la adaptación dinámica y no supervisada de modelos de lenguaje para el robustecimiento de un sistema automático de reconocimiento de voz. Esta adaptación toma como base distintos enfoque de los sistemas mencionados (de recuperación de información y aprendizaje de máquina) mediante los cuales buscamos identificar las temáticas sobre las cuales se está hablando en una grabación de audio. Dicha identificación, por lo tanto, permite realizar una adaptación del modelo de lenguaje de acuerdo a las condiciones del contexto. El marco de contextualización propuesto se puede dividir en dos sistemas principales: un sistema de identificación de temática y un sistema de adaptación dinámica de modelos de lenguaje. Esta Tesis puede describirse en detalle desde la perspectiva de las contribuciones particulares realizadas en cada uno de los campos que componen el marco propuesto: _ En lo referente al sistema de identificación de temática, nos hemos enfocado en aportar mejoras a las técnicas de pre-procesamiento de documentos, asimismo en contribuir a la definición de criterios más robustos para la selección de index-terms. – La eficiencia de los sistemas basados tanto en técnicas de recuperación de información como en técnicas de aprendizaje de máquina, y específicamente de aquellos sistemas que particularizan en la tarea de identificación de temática, depende, en gran medida, de los mecanismos de preprocesamiento que se aplican a los documentos. Entre las múltiples operaciones que hacen parte de un esquema de preprocesamiento, la selección adecuada de los términos de indexado (index-terms) es crucial para establecer relaciones semánticas y conceptuales entre los términos y los documentos. Este proceso también puede verse afectado, o bien por una mala elección de stopwords, o bien por la falta de precisión en la definición de reglas de lematización. En este sentido, en este trabajo comparamos y evaluamos diferentes criterios para el preprocesamiento de los documentos, así como también distintas estrategias para la selección de los index-terms. Esto nos permite no sólo reducir el tamaño de la estructura de indexación, sino también mejorar el proceso de identificación de temática. – Uno de los aspectos más importantes en cuanto al rendimiento de los sistemas de identificación de temática es la asignación de diferentes pesos a los términos de acuerdo a su contribución al contenido del documento. En este trabajo evaluamos y proponemos enfoques alternativos a los esquemas tradicionales de ponderado de términos (tales como tf-idf ) que nos permitan mejorar la especificidad de los términos, así como también discriminar mejor las temáticas de los documentos. _ Respecto a la adaptación dinámica de modelos de lenguaje, hemos dividimos el proceso de contextualización en varios pasos. – Para la generación de modelos de lenguaje basados en temática, proponemos dos tipos de enfoques: un enfoque supervisado y un enfoque no supervisado. En el primero de ellos nos basamos en las etiquetas de temática que originalmente acompañan a los documentos del corpus que empleamos. A partir de estas, agrupamos los documentos que forman parte de la misma temática y generamos modelos de lenguaje a partir de dichos grupos. Sin embargo, uno de los objetivos que se persigue en esta Tesis es evaluar si el uso de estas etiquetas para la generación de modelos es óptimo en términos del rendimiento del reconocedor. Por esta razón, nosotros proponemos un segundo enfoque, un enfoque no supervisado, en el cual el objetivo es agrupar, automáticamente, los documentos en clusters temáticos, basándonos en la similaridad semántica existente entre los documentos. Por medio de enfoques de agrupamiento conseguimos mejorar la cohesión conceptual y semántica en cada uno de los clusters, lo que a su vez nos permitió refinar los modelos de lenguaje basados en temática y mejorar el rendimiento del sistema de reconocimiento. – Desarrollamos diversas estrategias para generar un modelo de lenguaje dependiente del contexto. Nuestro objetivo es que este modelo refleje el contexto semántico del habla, i.e. las temáticas más relevantes que se están discutiendo. Este modelo es generado por medio de la interpolación lineal entre aquellos modelos de lenguaje basados en temática que estén relacionados con las temáticas más relevantes. La estimación de los pesos de interpolación está basada principalmente en el resultado del proceso de identificación de temática. – Finalmente, proponemos una metodología para la adaptación dinámica de un modelo de lenguaje general. El proceso de adaptación tiene en cuenta no sólo al modelo dependiente del contexto sino también a la información entregada por el proceso de identificación de temática. El esquema usado para la adaptación es una interpolación lineal entre el modelo general y el modelo dependiente de contexto. Estudiamos también diferentes enfoques para determinar los pesos de interpolación entre ambos modelos. Una vez definida la base teórica de nuestro marco de contextualización, proponemos su aplicación dentro de un sistema automático de reconocimiento de voz. Para esto, nos enfocamos en dos aspectos: la contextualización de los modelos de lenguaje empleados por el sistema y la incorporación de información semántica en el proceso de adaptación basado en temática. En esta Tesis proponemos un marco experimental basado en una arquitectura de reconocimiento en ‘dos etapas’. En la primera etapa, empleamos sistemas basados en técnicas de recuperación de información y aprendizaje de máquina para identificar las temáticas sobre las cuales se habla en una transcripción de un segmento de audio. Esta transcripción es generada por el sistema de reconocimiento empleando un modelo de lenguaje general. De acuerdo con la relevancia de las temáticas que han sido identificadas, se lleva a cabo la adaptación dinámica del modelo de lenguaje. En la segunda etapa de la arquitectura de reconocimiento, usamos este modelo adaptado para realizar de nuevo el reconocimiento del segmento de audio. Para determinar los beneficios del marco de trabajo propuesto, llevamos a cabo la evaluación de cada uno de los sistemas principales previamente mencionados. Esta evaluación es realizada sobre discursos en el dominio de la política usando la base de datos EPPS (European Parliamentary Plenary Sessions - Sesiones Plenarias del Parlamento Europeo) del proyecto europeo TC-STAR. Analizamos distintas métricas acerca del rendimiento de los sistemas y evaluamos las mejoras propuestas con respecto a los sistemas de referencia. ABSTRACT The last decade has witnessed major advances in speech recognition technology. Today’s commercial systems are able to recognize continuous speech from numerous speakers, with acceptable levels of error and without the need for an explicit adaptation procedure. Despite this progress, speech recognition is far from being a solved problem. Most of these systems are adjusted to a particular domain and their efficacy depends significantly, among many other aspects, on the similarity between the language model used and the task that is being addressed. This dependence is even more important in scenarios where the statistical properties of the language fluctuates throughout the time, for example, in application domains involving spontaneous and multitopic speech. Over the last years there has been an increasing effort in enhancing the speech recognition systems for such domains. This has been done, among other approaches, by means of techniques of automatic adaptation. These techniques are applied to the existing systems, specially since exporting the system to a new task or domain may be both time-consuming and expensive. Adaptation techniques require additional sources of information, and the spoken language could provide some of them. It must be considered that speech not only conveys a message, it also provides information on the context in which the spoken communication takes place (e.g. on the subject on which it is being talked about). Therefore, when we communicate through speech, it could be feasible to identify the elements of the language that characterize the context, and at the same time, to track the changes that occur in those elements over time. This information can be extracted and exploited through techniques of information retrieval and machine learning. This allows us, within the development of more robust speech recognition systems, to enhance the adaptation of language models to the conditions of the context, thus strengthening the recognition system for domains under changing conditions (such as potential variations in vocabulary, style and topic). In this sense, the main contribution of this Thesis is the proposal and evaluation of a framework of topic-motivated contextualization based on the dynamic and non-supervised adaptation of language models for the enhancement of an automatic speech recognition system. This adaptation is based on an combined approach (from the perspective of both information retrieval and machine learning fields) whereby we identify the topics that are being discussed in an audio recording. The topic identification, therefore, enables the system to perform an adaptation of the language model according to the contextual conditions. The proposed framework can be divided in two major systems: a topic identification system and a dynamic language model adaptation system. This Thesis can be outlined from the perspective of the particular contributions made in each of the fields that composes the proposed framework: _ Regarding the topic identification system, we have focused on the enhancement of the document preprocessing techniques in addition to contributing in the definition of more robust criteria for the selection of index-terms. – Within both information retrieval and machine learning based approaches, the efficiency of topic identification systems, depends, to a large extent, on the mechanisms of preprocessing applied to the documents. Among the many operations that encloses the preprocessing procedures, an adequate selection of index-terms is critical to establish conceptual and semantic relationships between terms and documents. This process might also be weakened by a poor choice of stopwords or lack of precision in defining stemming rules. In this regard we compare and evaluate different criteria for preprocessing the documents, as well as for improving the selection of the index-terms. This allows us to not only reduce the size of the indexing structure but also to strengthen the topic identification process. – One of the most crucial aspects, in relation to the performance of topic identification systems, is to assign different weights to different terms depending on their contribution to the content of the document. In this sense we evaluate and propose alternative approaches to traditional weighting schemes (such as tf-idf ) that allow us to improve the specificity of terms, and to better identify the topics that are related to documents. _ Regarding the dynamic language model adaptation, we divide the contextualization process into different steps. – We propose supervised and unsupervised approaches for the generation of topic-based language models. The first of them is intended to generate topic-based language models by grouping the documents, in the training set, according to the original topic labels of the corpus. Nevertheless, a goal of this Thesis is to evaluate whether or not the use of these labels to generate language models is optimal in terms of recognition accuracy. For this reason, we propose a second approach, an unsupervised one, in which the objective is to group the data in the training set into automatic topic clusters based on the semantic similarity between the documents. By means of clustering approaches we expect to obtain a more cohesive association of the documents that are related by similar concepts, thus improving the coverage of the topic-based language models and enhancing the performance of the recognition system. – We develop various strategies in order to create a context-dependent language model. Our aim is that this model reflects the semantic context of the current utterance, i.e. the most relevant topics that are being discussed. This model is generated by means of a linear interpolation between the topic-based language models related to the most relevant topics. The estimation of the interpolation weights is based mainly on the outcome of the topic identification process. – Finally, we propose a methodology for the dynamic adaptation of a background language model. The adaptation process takes into account the context-dependent model as well as the information provided by the topic identification process. The scheme used for the adaptation is a linear interpolation between the background model and the context-dependent one. We also study different approaches to determine the interpolation weights used in this adaptation scheme. Once we defined the basis of our topic-motivated contextualization framework, we propose its application into an automatic speech recognition system. We focus on two aspects: the contextualization of the language models used by the system, and the incorporation of semantic-related information into a topic-based adaptation process. To achieve this, we propose an experimental framework based in ‘a two stages’ recognition architecture. In the first stage of the architecture, Information Retrieval and Machine Learning techniques are used to identify the topics in a transcription of an audio segment. This transcription is generated by the recognition system using a background language model. According to the confidence on the topics that have been identified, the dynamic language model adaptation is carried out. In the second stage of the recognition architecture, an adapted language model is used to re-decode the utterance. To test the benefits of the proposed framework, we carry out the evaluation of each of the major systems aforementioned. The evaluation is conducted on speeches of political domain using the EPPS (European Parliamentary Plenary Sessions) database from the European TC-STAR project. We analyse several performance metrics that allow us to compare the improvements of the proposed systems against the baseline ones.