Feature selection in the reconstruction of complex network representations of spectral data

Complex networks have been extensively used in the last decade to characterize and analyze complex systems, and they have been recently proposed as a novel instrument for the analysis of spectra extracted from biological samples. Yet, the high number of measurements composing spectra, and the consequent high computational cost, make a direct network analysis unfeasible. We here present a comparative analysis of three customary feature selection algorithms, including the binning of spectral data and the use of information theory metrics. Such algorithms are compared by assessing the score obtained in a classification task, where healthy subjects and people suffering from different types of cancers should be discriminated. Results indicate that a feature selection strategy based on Mutual Information outperforms the more classical data binning, while allowing a reduction of the dimensionality of the data set in two orders of magnitude

Metodologías de modelado, monitorización y asistencia robótica en neurorrehabilitación funcional de extremidad superior

La Organización Mundial de la Salud (OMS) prevé que para el año 2020, el Daño Cerebral Adquirido (DCA) estará entre las 10 causas más comunes de discapacidad. Estas lesiones, dadas sus consecuencias físicas, sensoriales, cognitivas, emocionales y socioeconómicas, cambian dramáticamente la vida de los pacientes y sus familias. Las nuevas técnicas de intervención precoz y el desarrollo de la medicina intensiva en la atención al DCA han mejorado notablemente la probabilidad de supervivencia. Sin embargo, hoy por hoy, las lesiones cerebrales no tienen ningún tratamiento quirúrgico que tenga por objetivo restablecer la funcionalidad perdida, sino que las terapias rehabilitadoras se dirigen hacia la compensación de los déficits producidos. Uno de los objetivos principales de la neurorrehabilitación es, por tanto, dotar al paciente de la capacidad necesaria para ejecutar las Actividades de Vida Diaria (AVDs) necesarias para desarrollar una vida independiente, siendo fundamentales aquellas en las que la Extremidad Superior (ES) está directamente implicada, dada su gran importancia a la hora de la manipulación de objetos. Con la incorporación de nuevas soluciones tecnológicas al proceso de neurorrehabilitación se pretende alcanzar un nuevo paradigma centrado en ofrecer una práctica personalizada, monitorizada y ubicua con una valoración continua de la eficacia y de la eficiencia de los procedimientos y con capacidad de generar conocimientos que impulsen la ruptura del paradigma de actual. Los nuevos objetivos consistirán en minimizar el impacto de las enfermedades que afectan a la capacidad funcional de las personas, disminuir el tiempo de incapacidad y permitir una gestión más eficiente de los recursos. Estos objetivos clínicos, de gran impacto socio-económico, sólo pueden alcanzarse desde una apuesta decidida en nuevas tecnologías, metodologías y algoritmos capaces de ocasionar la ruptura tecnológica necesaria que permita superar las barreras que hasta el momento han impedido la penetración tecnológica en el campo de la rehabilitación de manera universal. De esta forma, los trabajos y resultados alcanzados en la Tesis son los siguientes: 1. Modelado de AVDs: como paso previo a la incorporación de ayudas tecnológicas al proceso rehabilitador, se hace necesaria una primera fase de modelado y formalización del conocimiento asociado a la ejecución de las actividades que se realizan como parte de la terapia. En particular, las tareas más complejas y a su vez con mayor repercusión terapéutica son las AVDs, cuya formalización permitirá disponer de modelos de movimiento sanos que actuarán de referencia para futuros desarrollos tecnológicos dirigidos a personas con DCA. Siguiendo una metodología basada en diagramas de estados UML se han modelado las AVDs 'servir agua de una jarra' y 'coger un botella' 2. Monitorización ubícua del movimiento de la ES: se ha diseñado, desarrollado y validado un sistema de adquisición de movimiento basado en tecnología inercial que mejora las limitaciones de los dispositivos comerciales actuales (coste muy elevado e incapacidad para trabajar en entornos no controlados); los altos coeficientes de correlación y los bajos niveles de error obtenidos en los corregistros llevados a cabo con el sistema comercial BTS SMART-D demuestran la alta precisión del sistema. También se ha realizado un trabajo de investigación exploratorio de un sistema de captura de movimiento de coste muy reducido basado en visión estereoscópica, habiéndose detectado los puntos clave donde se hace necesario incidir desde un punto de vista tecnológico para su incorporación en un entorno real 3. Resolución del Problema Cinemático Inverso (PCI): se ha diseñado, desarrollado y validado una solución al PCI cuando el manipulador se corresponde con una ES humana estudiándose 2 posibles alternativas, una basada en la utilización de un Perceptrón Multicapa (PMC) y otra basada en sistemas Artificial Neuro-Fuzzy Inference Systems (ANFIS). La validación, llevada a cabo utilizando información relativa a los modelos disponibles de AVDs, indica que una solución basada en un PMC con 3 neuronas en la capa de entrada, una capa oculta también de 3 neuronas y una capa de salida con tantas neuronas como Grados de Libertad (GdLs) tenga el modelo de la ES, proporciona resultados, tanto de precisión como de tiempo de cálculo, que la hacen idónea para trabajar en sistemas con requisitos de tiempo real 4. Control inteligente assisted-as-needed: se ha diseñado, desarrollado y validado un algoritmo de control assisted-as-needed para una ortesis robótica con capacidades de actuación anticipatoria de la que existe un prototipo implementado en la actualidad. Los resultados obtenidos demuestran cómo el sistema es capaz de adaptarse al perfil disfuncional del paciente activando la ayuda en instantes anteriores a la ocurrencia de movimientos incorrectos. Esta estrategia implica un aumento en la participación del paciente y, por tanto, en su actividad muscular, fomentándose los procesos la plasticidad cerebral responsables del reaprendizaje o readaptación motora 5. Simuladores robóticos para planificación: se propone la utilización de un simulador robótico assisted-as-needed como herramienta de planificación de sesiones de rehabilitación personalizadas y con un objetivo clínico marcado en las que interviene una ortesis robotizada. Los resultados obtenidos evidencian como, tras la ejecución de ciertos algoritmos sencillos, es posible seleccionar automáticamente una configuración para el algoritmo de control assisted-as-needed que consigue que la ortesis se adapte a los criterios establecidos desde un punto de vista clínico en función del paciente estudiado. Estos resultados invitan a profundizar en el desarrollo de algoritmos más avanzados de selección de parámetros a partir de baterías de simulaciones Estos trabajos han servido para corroborar las hipótesis de investigación planteadas al inicio de la misma, permitiendo, asimismo, la apertura de nuevas líneas de investigación. Summary The World Health Organization (WHO) predicts that by the year 2020, Acquired Brain Injury (ABI) will be among the ten most common ailments. These injuries dramatically change the life of the patients and their families due to their physical, sensory, cognitive, emotional and socio-economic consequences. New techniques of early intervention and the development of intensive ABI care have noticeably improved the survival rate. However, in spite of these advances, brain injuries still have no surgical or pharmacological treatment to re-establish the lost functions. Neurorehabilitation therapies address this problem by restoring, minimizing or compensating the functional alterations in a person disabled because of a nervous system injury. One of the main objectives of Neurorehabilitation is to provide patients with the capacity to perform specific Activities of the Daily Life (ADL) required for an independent life, especially those in which the Upper Limb (UL) is directly involved due to its great importance in manipulating objects within the patients' environment. The incorporation of new technological aids to the neurorehabilitation process tries to reach a new paradigm focused on offering a personalized, monitored and ubiquitous practise with continuous assessment of both the efficacy and the efficiency of the procedures and with the capacity of generating new knowledge. New targets will be to minimize the impact of the sicknesses affecting the functional capabilitiies of the subjects, to decrease the time of the physical handicap and to allow a more efficient resources handling. These targets, of a great socio-economic impact, can only be achieved by means of new technologies and algorithms able to provoke the technological break needed to beat the barriers that are stopping the universal penetration of the technology in the field of rehabilitation. In this way, this PhD Thesis has achieved the following results: 1. ADL Modeling: as a previous step to the incorporation of technological aids to the neurorehabilitation process, it is necessary a first modelling and formalization phase of the knowledge associated to the execution of the activities that are performed as a part of the therapy. In particular, the most complex and therapeutically relevant tasks are the ADLs, whose formalization will produce healthy motion models to be used as a reference for future technological developments. Following a methodology based on UML state-chart diagrams, the ADLs 'serving water from a jar' and 'picking up a bottle' have been modelled 2. Ubiquitous monitoring of the UL movement: it has been designed, developed and validated a motion acquisition system based on inertial technology that improves the limitations of the current devices (high monetary cost and inability of working within uncontrolled environments); the high correlation coefficients and the low error levels obtained throughout several co-registration sessions with the commercial sys- tem BTS SMART-D show the high precision of the system. Besides an exploration of a very low cost stereoscopic vision-based motion capture system has been carried out and the key points where it is necessary to insist from a technological point of view have been detected 3. Inverse Kinematics (IK) problem solving: a solution to the IK problem has been proposed for a manipulator that corresponds to a human UL. This solution has been faced by means of two different alternatives, one based on a Mulilayer Perceptron (MLP) and another based on Artificial Neuro-Fuzzy Inference Systems (ANFIS). The validation of these solutions, carried out using the information regarding the previously generated motion models, indicate that a MLP-based solution, with an architecture consisting in 3 neurons in the input layer, one hidden layer of 3 neurons and an output layer with as many neurons as the number of Degrees of Freedom (DoFs) that the UL model has, is the one that provides the best results both in terms of precission and in terms of processing time, making in idoneous to be integrated within a system with real time restrictions 4. Assisted-as-needed intelligent control: an assisted-as-needed control algorithm with anticipatory actuation capabilities has been designed, developed and validated for a robotic orthosis of which there is an already implemented prototype. Obtained results demonstrate that the control system is able to adapt to the dysfunctional profile of the patient by triggering the assistance right before an incorrect movement is going to take place. This strategy implies an increase in the participation of the patients and in his or her muscle activity, encouraging the neural plasticity processes in charge of the motor learning 5. Planification with a robotic simulator: in this work a robotic simulator is proposed as a planification tool for personalized rehabilitation sessions under a certain clinical criterium. Obtained results indicate that, after the execution of simple parameter selection algorithms, it is possible to automatically choose a specific configuration that makes the assisted-as-needed control algorithm to adapt both to the clinical criteria and to the patient. These results invite researchers to work in the development of more complex parameter selection algorithms departing from simulation batteries Obtained results have been useful to corroborate the hypotheses set out at the beginning of this PhD Thesis. Besides, they have allowed the creation of new research lines in all the studied application fields.

Comparison of metaheuristic strategies for peakbin selection in proteomic mass spectrometry data

Mass spectrometry (MS) data provide a promising strategy for biomarker discovery. For this purpose, the detection of relevant peakbins in MS data is currently under intense research. Data from mass spectrometry are challenging to analyze because of their high dimensionality and the generally low number of samples available. To tackle this problem, the scientific community is becoming increasingly interested in applying feature subset selection techniques based on specialized machine learning algorithms. In this paper, we present a performance comparison of some metaheuristics: best first (BF), genetic algorithm (GA), scatter search (SS) and variable neighborhood search (VNS). Up to now, all the algorithms, except for GA, have been first applied to detect relevant peakbins in MS data. All these metaheuristic searches are embedded in two different filter and wrapper schemes coupled with Naive Bayes and SVM classifiers.

Response threshold models, stochastic learning automata and ant colony optimization-based decentralized self-coordination algorithms for heterogeneous multi-tasks distribution in multi-robot systems

In recent decades, there has been an increasing interest in systems comprised of several autonomous mobile robots, and as a result, there has been a substantial amount of development in the eld of Articial Intelligence, especially in Robotics. There are several studies in the literature by some researchers from the scientic community that focus on the creation of intelligent machines and devices capable to imitate the functions and movements of living beings. Multi-Robot Systems (MRS) can often deal with tasks that are dicult, if not impossible, to be accomplished by a single robot. In the context of MRS, one of the main challenges is the need to control, coordinate and synchronize the operation of multiple robots to perform a specic task. This requires the development of new strategies and methods which allow us to obtain the desired system behavior in a formal and concise way. This PhD thesis aims to study the coordination of multi-robot systems, in particular, addresses the problem of the distribution of heterogeneous multi-tasks. The main interest in these systems is to understand how from simple rules inspired by the division of labor in social insects, a group of robots can perform tasks in an organized and coordinated way. We are mainly interested on truly distributed or decentralized solutions in which the robots themselves, autonomously and in an individual manner, select a particular task so that all tasks are optimally distributed. In general, to perform the multi-tasks distribution among a team of robots, they have to synchronize their actions and exchange information. Under this approach we can speak of multi-tasks selection instead of multi-tasks assignment, which means, that the agents or robots select the tasks instead of being assigned a task by a central controller. The key element in these algorithms is the estimation ix of the stimuli and the adaptive update of the thresholds. This means that each robot performs this estimate locally depending on the load or the number of pending tasks to be performed. In addition, it is very interesting the evaluation of the results in function in each approach, comparing the results obtained by the introducing noise in the number of pending loads, with the purpose of simulate the robot's error in estimating the real number of pending tasks. The main contribution of this thesis can be found in the approach based on self-organization and division of labor in social insects. An experimental scenario for the coordination problem among multiple robots, the robustness of the approaches and the generation of dynamic tasks have been presented and discussed. The particular issues studied are: Threshold models: It presents the experiments conducted to test the response threshold model with the objective to analyze the system performance index, for the problem of the distribution of heterogeneous multitasks in multi-robot systems; also has been introduced additive noise in the number of pending loads and has been generated dynamic tasks over time. Learning automata methods: It describes the experiments to test the learning automata-based probabilistic algorithms. The approach was tested to evaluate the system performance index with additive noise and with dynamic tasks generation for the same problem of the distribution of heterogeneous multi-tasks in multi-robot systems. Ant colony optimization: The goal of the experiments presented is to test the ant colony optimization-based deterministic algorithms, to achieve the distribution of heterogeneous multi-tasks in multi-robot systems. In the experiments performed, the system performance index is evaluated by introducing additive noise and dynamic tasks generation over time.

Regularized logistic regression and multi-objective variable selection for classifying MEG data

This paper addresses the question of maximizing classifier accuracy for classifying task-related mental activity from Magnetoencelophalography (MEG) data. We propose the use of different sources of information and introduce an automatic channel selection procedure. To determine an informative set of channels, our approach combines a variety of machine learning algorithms: feature subset selection methods, classifiers based on regularized logistic regression, information fusion, and multiobjective optimization based on probabilistic modeling of the search space. The experimental results show that our proposal is able to improve classification accuracy compared to approaches whose classifiers use only one type of MEG information or for which the set of channels is fixed a priori.

Automatic blood glucose classification for gestational diabetes with feature selection: decision trees vs neural networks

Automatic blood glucose classification may help specialists to provide a better interpretation of blood glucose data, downloaded directly from patients glucose meter and will contribute in the development of decision support systems for gestational diabetes. This paper presents an automatic blood glucose classifier for gestational diabetes that compares 6 different feature selection methods for two machine learning algorithms: neural networks and decision trees. Three searching algorithms, Greedy, Best First and Genetic, were combined with two different evaluators, CSF and Wrapper, for the feature selection. The study has been made with 6080 blood glucose measurements from 25 patients. Decision trees with a feature set selected with the Wrapper evaluator and the Best first search algorithm obtained the best accuracy: 95.92%.

Cognitive reserve is associated with the functional organization of brain networks in healthy aging: a MEG study

The proportion of elderly people in the population has increased rapidly in the last century and consequently "healthy aging" is expected to become a critical area of research in neuroscience. Evidence reveals how healthy aging depends on three main behavioral factors: social lifestyle, cognitive activity and physical activity. In this study, we focused on the role of cognitive activity, concentrating specifically on educational and occupational attainment factors, which were considered two of the main pillars of cognitive reserve. 21 subjects with similar rates of social lifestyle, physical and cognitive activity were selected from a sample of 55 healthy adults. These subjects were divided into two groups according to their level of cognitive reserve; one group comprised subjects with high cognitive reserve (9 members) and the other contained those with low cognitive reserve (12 members). To evaluate the cortical brain connectivity network, all participants were recorded by Magnetoencephalography (MEG) while they performed a memory task (modified version of the Sternberg¿s Task). We then applied two algorithms (Phase Locking Value & Phase-Lag Index) to study the dynamics of functional connectivity. In response to the same task, the subjects with lower cognitive reserve presented higher functional connectivity than those with higher cognitive reserve. These results may indicate that participants with low cognitive reserve needed a greater 'effort' than those with high cognitive reserve to achieve the same level of cognitive performance. Therefore, we conclude that cognitive reserve contributes to the modulation of the functional connectivity patterns of the aging brain.

Parallel Computer Vision Algorithms for Graphics Processing Units

La evolución de los teléfonos móviles inteligentes, dotados de cámaras digitales, está provocando una creciente demanda de aplicaciones cada vez más complejas que necesitan algoritmos de visión artificial en tiempo real; puesto que el tamaño de las señales de vídeo no hace sino aumentar y en cambio el rendimiento de los procesadores de un solo núcleo se ha estancado, los nuevos algoritmos que se diseñen para visión artificial han de ser paralelos para poder ejecutarse en múltiples procesadores y ser computacionalmente escalables. Una de las clases de procesadores más interesantes en la actualidad se encuentra en las tarjetas gráficas (GPU), que son dispositivos que ofrecen un alto grado de paralelismo, un excelente rendimiento numérico y una creciente versatilidad, lo que los hace interesantes para llevar a cabo computación científica. En esta tesis se exploran dos aplicaciones de visión artificial que revisten una gran complejidad computacional y no pueden ser ejecutadas en tiempo real empleando procesadores tradicionales. En cambio, como se demuestra en esta tesis, la paralelización de las distintas subtareas y su implementación sobre una GPU arrojan los resultados deseados de ejecución con tasas de refresco interactivas. Asimismo, se propone una técnica para la evaluación rápida de funciones de complejidad arbitraria especialmente indicada para su uso en una GPU. En primer lugar se estudia la aplicación de técnicas de síntesis de imágenes virtuales a partir de únicamente dos cámaras lejanas y no paralelas—en contraste con la configuración habitual en TV 3D de cámaras cercanas y paralelas—con información de color y profundidad. Empleando filtros de mediana modificados para la elaboración de un mapa de profundidad virtual y proyecciones inversas, se comprueba que estas técnicas son adecuadas para una libre elección del punto de vista. Además, se demuestra que la codificación de la información de profundidad con respecto a un sistema de referencia global es sumamente perjudicial y debería ser evitada. Por otro lado se propone un sistema de detección de objetos móviles basado en técnicas de estimación de densidad con funciones locales. Este tipo de técnicas es muy adecuada para el modelado de escenas complejas con fondos multimodales, pero ha recibido poco uso debido a su gran complejidad computacional. El sistema propuesto, implementado en tiempo real sobre una GPU, incluye propuestas para la estimación dinámica de los anchos de banda de las funciones locales, actualización selectiva del modelo de fondo, actualización de la posición de las muestras de referencia del modelo de primer plano empleando un filtro de partículas multirregión y selección automática de regiones de interés para reducir el coste computacional. Los resultados, evaluados sobre diversas bases de datos y comparados con otros algoritmos del estado del arte, demuestran la gran versatilidad y calidad de la propuesta. Finalmente se propone un método para la aproximación de funciones arbitrarias empleando funciones continuas lineales a tramos, especialmente indicada para su implementación en una GPU mediante el uso de las unidades de filtraje de texturas, normalmente no utilizadas para cómputo numérico. La propuesta incluye un riguroso análisis matemático del error cometido en la aproximación en función del número de muestras empleadas, así como un método para la obtención de una partición cuasióptima del dominio de la función para minimizar el error. ABSTRACT The evolution of smartphones, all equipped with digital cameras, is driving a growing demand for ever more complex applications that need to rely on real-time computer vision algorithms. However, video signals are only increasing in size, whereas the performance of single-core processors has somewhat stagnated in the past few years. Consequently, new computer vision algorithms will need to be parallel to run on multiple processors and be computationally scalable. One of the most promising classes of processors nowadays can be found in graphics processing units (GPU). These are devices offering a high parallelism degree, excellent numerical performance and increasing versatility, which makes them interesting to run scientific computations. In this thesis, we explore two computer vision applications with a high computational complexity that precludes them from running in real time on traditional uniprocessors. However, we show that by parallelizing subtasks and implementing them on a GPU, both applications attain their goals of running at interactive frame rates. In addition, we propose a technique for fast evaluation of arbitrarily complex functions, specially designed for GPU implementation. First, we explore the application of depth-image–based rendering techniques to the unusual configuration of two convergent, wide baseline cameras, in contrast to the usual configuration used in 3D TV, which are narrow baseline, parallel cameras. By using a backward mapping approach with a depth inpainting scheme based on median filters, we show that these techniques are adequate for free viewpoint video applications. In addition, we show that referring depth information to a global reference system is ill-advised and should be avoided. Then, we propose a background subtraction system based on kernel density estimation techniques. These techniques are very adequate for modelling complex scenes featuring multimodal backgrounds, but have not been so popular due to their huge computational and memory complexity. The proposed system, implemented in real time on a GPU, features novel proposals for dynamic kernel bandwidth estimation for the background model, selective update of the background model, update of the position of reference samples of the foreground model using a multi-region particle filter, and automatic selection of regions of interest to reduce computational cost. The results, evaluated on several databases and compared to other state-of-the-art algorithms, demonstrate the high quality and versatility of our proposal. Finally, we propose a general method for the approximation of arbitrarily complex functions using continuous piecewise linear functions, specially formulated for GPU implementation by leveraging their texture filtering units, normally unused for numerical computation. Our proposal features a rigorous mathematical analysis of the approximation error in function of the number of samples, as well as a method to obtain a suboptimal partition of the domain of the function to minimize approximation error.