New definitions of margin for multi-class Boosting algorithms

La familia de algoritmos de Boosting son un tipo de técnicas de clasificación y regresión que han demostrado ser muy eficaces en problemas de Visión Computacional. Tal es el caso de los problemas de detección, de seguimiento o bien de reconocimiento de caras, personas, objetos deformables y acciones. El primer y más popular algoritmo de Boosting, AdaBoost, fue concebido para problemas binarios. Desde entonces, muchas han sido las propuestas que han aparecido con objeto de trasladarlo a otros dominios más generales: multiclase, multilabel, con costes, etc. Nuestro interés se centra en extender AdaBoost al terreno de la clasificación multiclase, considerándolo como un primer paso para posteriores ampliaciones. En la presente tesis proponemos dos algoritmos de Boosting para problemas multiclase basados en nuevas derivaciones del concepto margen. El primero de ellos, PIBoost, está concebido para abordar el problema descomponiéndolo en subproblemas binarios. Por un lado, usamos una codificación vectorial para representar etiquetas y, por otro, utilizamos la función de pérdida exponencial multiclase para evaluar las respuestas. Esta codificación produce un conjunto de valores margen que conllevan un rango de penalizaciones en caso de fallo y recompensas en caso de acierto. La optimización iterativa del modelo genera un proceso de Boosting asimétrico cuyos costes dependen del número de etiquetas separadas por cada clasificador débil. De este modo nuestro algoritmo de Boosting tiene en cuenta el desbalanceo debido a las clases a la hora de construir el clasificador. El resultado es un método bien fundamentado que extiende de manera canónica al AdaBoost original. El segundo algoritmo propuesto, BAdaCost, está concebido para problemas multiclase dotados de una matriz de costes. Motivados por los escasos trabajos dedicados a generalizar AdaBoost al terreno multiclase con costes, hemos propuesto un nuevo concepto de margen que, a su vez, permite derivar una función de pérdida adecuada para evaluar costes. Consideramos nuestro algoritmo como la extensión más canónica de AdaBoost para este tipo de problemas, ya que generaliza a los algoritmos SAMME, Cost-Sensitive AdaBoost y PIBoost. Por otro lado, sugerimos un simple procedimiento para calcular matrices de coste adecuadas para mejorar el rendimiento de Boosting a la hora de abordar problemas estándar y problemas con datos desbalanceados. Una serie de experimentos nos sirven para demostrar la efectividad de ambos métodos frente a otros conocidos algoritmos de Boosting multiclase en sus respectivas áreas. En dichos experimentos se usan bases de datos de referencia en el área de Machine Learning, en primer lugar para minimizar errores y en segundo lugar para minimizar costes. Además, hemos podido aplicar BAdaCost con éxito a un proceso de segmentación, un caso particular de problema con datos desbalanceados. Concluimos justificando el horizonte de futuro que encierra el marco de trabajo que presentamos, tanto por su aplicabilidad como por su flexibilidad teórica. Abstract The family of Boosting algorithms represents a type of classification and regression approach that has shown to be very effective in Computer Vision problems. Such is the case of detection, tracking and recognition of faces, people, deformable objects and actions. The first and most popular algorithm, AdaBoost, was introduced in the context of binary classification. Since then, many works have been proposed to extend it to the more general multi-class, multi-label, costsensitive, etc... domains. Our interest is centered in extending AdaBoost to two problems in the multi-class field, considering it a first step for upcoming generalizations. In this dissertation we propose two Boosting algorithms for multi-class classification based on new generalizations of the concept of margin. The first of them, PIBoost, is conceived to tackle the multi-class problem by solving many binary sub-problems. We use a vectorial codification to represent class labels and a multi-class exponential loss function to evaluate classifier responses. This representation produces a set of margin values that provide a range of penalties for failures and rewards for successes. The stagewise optimization of this model introduces an asymmetric Boosting procedure whose costs depend on the number of classes separated by each weak-learner. In this way the Boosting procedure takes into account class imbalances when building the ensemble. The resulting algorithm is a well grounded method that canonically extends the original AdaBoost. The second algorithm proposed, BAdaCost, is conceived for multi-class problems endowed with a cost matrix. Motivated by the few cost-sensitive extensions of AdaBoost to the multi-class field, we propose a new margin that, in turn, yields a new loss function appropriate for evaluating costs. Since BAdaCost generalizes SAMME, Cost-Sensitive AdaBoost and PIBoost algorithms, we consider our algorithm as a canonical extension of AdaBoost to this kind of problems. We additionally suggest a simple procedure to compute cost matrices that improve the performance of Boosting in standard and unbalanced problems. A set of experiments is carried out to demonstrate the effectiveness of both methods against other relevant Boosting algorithms in their respective areas. In the experiments we resort to benchmark data sets used in the Machine Learning community, firstly for minimizing classification errors and secondly for minimizing costs. In addition, we successfully applied BAdaCost to a segmentation task, a particular problem in presence of imbalanced data. We conclude the thesis justifying the horizon of future improvements encompassed in our framework, due to its applicability and theoretical flexibility.

Dinámicas de ocupación urbana del Anillo Verde metropolitano, desde sus orígenes en el Plan General del Área Metropolitana de Madrid de 1963 hasta el Plan General de Ordenación Urbana de 1997

El Anillo Verde metropolitano, definido por el Plan General de Ordenación Urbana del Área Metropolitana de Madrid en 1963 siguiendo los modelos planteados por la cultura urbanística internacional, como armadura de la estructura urbana del AMM, espacio protagonista dentro del sistema de espacios libres, lugar de uso público destinado al recreo y contacto con la naturaleza de la población madrileña, se convierte en realidad en una reserva de suelo que va entrando en juego motivado por las alianzas entre el poder institucional y la clase social dominante actuando al margen del planeamiento, poniendo en evidencia la escasez de recursos legales y culturales disponibles para la salvaguarda de los intereses comunes y, donde los condicionantes geográficos y naturales del territorio madrileño han influido decisivamente en la especialización funcional y espacial del Área Metropolitana de Madrid. Así pues considerando esta idea como HIPÓTESIS, el objetivo de la TESIS sería demostrarla, para lo cual se hace necesario primero, acotar espacial y temporalmente el objeto de estudio, es decir, del Anillo Verde metropolitano1, segundo, contextualizar histórica y disciplinarmente los presupuestos teóricos que conformaban la idea del Anillo Verde, tercero, reconocer, localizar y documentar las piezas que han ido materializando la ocupación urbana del Anillo Verde, clasificándolas según parámetros temporales, funcionales, urbanísticos y, formales, lo que permite analizar la geografía, uso, instrumentación y forma de su transformación a escala general metropolitana y, cuarto, profundizar a modo de comprobación a escala municipal y urbana en dos escenarios representativos del conjunto metropolitano: el municipio de Pozuelo de Alarcón y el distrito de Hortaleza-Barajas. El contenido del documento se divide en tres bloques, el bloque I, se centra en las bases teóricas, el bloque II sitúa el hilo argumental de la tesis a escala metropolitana y el bloque III comprueba el fenómeno a escala municipal y urbana. De esta forma, se comienza por la comprensión del significado del concepto del Anillo Verde, que va más allá de la dimensión instrumental asignada de límite y contención urbana frente al crecimiento de la ciudad industrial de principios del siglo XX, basada en la descentralización de la ciudad tradicional, para adquirir un significado más complejo, como gran espacio de reserva y salvaguarda de valores naturales y culturales que se expresaban en su territorio y que permitirían alcanzar el equilibrio entre la ciudad y sus habitantes, es decir, entre el hombre y el espacio que habita. Se hace un recorrido por las principales corrientes urbanísticas que se van nutriendo de distintas disciplinas (economía, sociología, geografía, biología, ecología) para plantear teorías que permitieran materializar un nuevo orden urbano según principios de equidad social, económica y ambiental, en una secuencia donde Europa y Estados Unidos realizaban un constante intercambio -el movimiento de la Ciudad Jardín o el Regionalismo, que dieron paso a propuestas como el Greater London o el Gran Berlín, donde la figura del Anillo Verde tenía un papel protagonista, y del que también participaría nuestro país y la ciudad de Madrid, con modelos regionales como el Plan Besteiro y urbanos como el Plan Bidagor, antecedentes directos del Plan General de Ordenación Urbana del Área Metropolitana de Madrid de 1963 que pone en marcha la ordenación del crecimiento metropolitano de Madrid. El hilo argumental de la tesis se organiza en una doble aproximación: un acercamiento a escala metropolitana a partir del reconocimiento del modelo de ciudad definido en los distintos planes generales que acompañaron el desarrollo metropolitano (municipio de Madrid y de los siete términos municipales que rodeaban a este y que tenían suelo destinado a Anillo Verde), haciendo referencia además a las relaciones con el planeamiento regional, concretando en una escala de aproximación municipal que avanza hasta la interpretación urbana detallada. El primer acercamiento tiene lugar en el bloque II y se organiza en tres capítulos. El capítulo 4 se dedica al punto obligado de partida de la geografía local, describiendo las características biofísicas de los terrenos que formaban parte del Anillo Verde, que han marcado históricamente la forma de aprovechamiento del territorio, desde las extensiones de bosques mediterráneos al norte y al oeste continuación del Monte del Pardo, a los distintos tipos de cultivo que se adaptaban al sustrato geológico y la forma del terreno (de las suaves ondulaciones de sedimentos arcósicos al norte a las extensas plataformas arenosas y yesíferas del sur), además de las zonas de huertos aprovechando las depresiones y los cursos de agua (arroyo del Monte Carmelo, arroyo de Valdebebas, arroyo del Quinto, arroyo del Santo, arroyo Butarque, arroyo Meaques y arroyo Pozuelo). Una vez reconocida la realidad física, el capítulo 5, avanza en la descripción de los distintos modelos de ciudad propuestos desde el planeamiento urbanístico, en sus distintas escalas, la regional y la municipal, como respuesta a la coyuntura social, económica y política que ha caracterizado el proceso de ocupación del Anillo Verde al compás de la construcción del AMM. Se han reunido las propuestas de planeamiento municipal de los distintos municipios que disponían de terreno calificado como Anillo Verde: Madrid, Coslada, Getafe, Leganés, Alcorcón, Boadilla del Monte y Pozuelo de Alarcón. Además se han incorporado las distintas propuestas de ordenación territorial que han servido de referencia al planeamiento municipal, en todas sus versiones, desde las sectoriales, de mayor éxito y apoyo institucional, a los distintos intentos de ordenación integral, de mayor complejidad pero de menor calado, precisamente por la dificultad de consenso entre la ordenación física y el desarrollo económico, entre los intereses privados y el beneficio público. El primer horizonte, comienza con la formulación del Plan General de Ordenación Urbana del Área Metropolitana de Madrid de 1963, su desarrollo y la puesta en marcha de los primeros planes municipales en la década de los años setenta, donde se comprueba la necesidad de un marco regional que “ordene” el territorio de forma integral y sirva de referencia a las actuaciones sectoriales que habían marcado el primer desarrollo metropolitano. El segundo, se sitúa dos décadas más tarde con la aprobación del Plan General de Ordenación Urbana de Madrid de 1985 y el conjunto de planes municipales de los términos limítrofes, que siguen su filosofía de austeridad en cuanto a crecimiento territorial. El tercero se inicia en 1997 con la siguiente generación de planes de corte neoliberal que imponen un modelo territorial basado en las grandes operaciones metropolitanas de centralidad, infraestructuras y equipamiento, que consumen de forma indiscriminada la totalidad del territorio madrileño. Será en el último capítulo del segundo bloque (capítulo 6) donde se represente gráficamente a escala metropolitana y se analicen las 229 piezas que han ido colmatando el espacio destinado a Anillo Verde, según los parámetros de estudio, en base a las cuales se plantean las primeras conclusiones generales de la tesis, poniendo de manifiesto que las alianzas entre los agentes soberanos en la construcción de la ciudad y su entorno han trasgredido sucesivamente las determinaciones del Planeamiento en su definición de modelo de ciudad y territorio, acusando la carencia de recursos instrumentales y jurídicos que alentaron el proceso de su desmantelamiento, y revelando la influencia de los condicionantes geográficos y naturales en la especialización funcional y segregación social en el conjunto del Área Metropolitana de Madrid. Se remata el discurso metropolitano con una batería de conclusiones que interpretan el fenómeno de ocupación del anillo de verdor metropolitano confirmando las hipótesis iniciales, reconociendo los valores medioambientales y culturales trasgredidos, sus diversos actores, las numerosas operaciones urbanísticas desarrolladas con distintos usos y envergadura, así como los instrumentos de planeamiento utilizados, en base a las cuales se materializa la construcción del AMM según un modelo extendido (spread), dibujando una mancha de aceite (o grase-spots según Geddes) que precisamente había querido evitarse desde el planeamiento urbanístico con la definición de un Anillo Verde, espacio inmune a la edificación, que se aleja de su papel estructurante (equilibrador entre la ciudad y sus habitantes) para convertirse en armadura de la estructura comunicativa, que una vez consolidada se convierte en la mejor aliada de la máquina inmobiliaria. El último paso, se desarrolla en el bloque III que se divide en los capítulo 7,8 y 9 y supone la comprobación de lo descrito en el conjunto de escala metropolitana, en dos aspectos fundamentales, la falta de consideración por los valores culturales y medioambientales que han modelado el territorio, imprimiéndole un carácter singular y específico y, la estructura del dominio del suelo, donde se reconoce de forma precisa el grupo social y los agentes encargados en cada momento de comercializar los suelos del anillo, que bajo el paraguas de la urgencia social y el engañoso beneficio popular, obtienen importantes beneficios económicos. Con esa intención, se da un salto hacia la escala municipal y urbana, seleccionando dos escenarios de estudio, el municipio de Pozuelo de Alarcón, que representa la materialización del crecimiento suburbano de la élite madrileña ocupando las zonas de mayor valor ecológico del anillo, y el distrito de Hortaleza-Barajas que ofrece su territorio a las grandes operaciones metropolitanas, apoyándose en el eje de actividad marcado por la conexión Madrid-Barcelona y el sistema aeroportuario de escala global, ambos situados al norte de la línea de borde entre la Sierra y la Mancha, ocupando por tanto los lugares más valiosos de la geografía madrileña (estructura funcional anticipada por Bidagor en 1946 en su modelo de ciudad adaptada al territorio madrileño) Una vez descrito este proceso trasgresor de límites, de normas, de conductas, y desde una perspectiva del fenómeno suficientemente documentada, en el capítulo 10, se realiza una reflexión sobre la incidencia real de la propuesta urbanística del Anillo Verde en la construcción del AMM, de la misma forma que se sugieren nuevos roles al planeamiento en un formato intencionado de largo recorrido en oposición a lo inmediato y circunstancial, que permita hacer una nueva lectura de los presupuestos teóricos que conformaban la idea del Anillo Verde, espacio articulador (medioambiental, social y cultural) del territorio madrileño. ABSTRACT The Metropolitan Greenbelt was defined by the 1963 Master Plan for the Madrid Metropolitan Area (MMA), following established international models of urban development, as the structural framework of the MMA, the principal open space within its network of open spaces and a public area of recreation and contact with nature for the residents of Madrid. In reality, however, it ha become a reserve of land in which various alliances between the institutional authorities and the dominant social class have been operating on the margin of the original plan, exposing a scarcity of legal and cultural resources for the safeguarding of common interests, and in which the geographical and natural characteristics of the territory itself have come to play an influential role in the functional specialization and spatial segregation of the MMA. With that idea as its HYPOTHESIS, the aim of this THESIS is to demonstrate its reality. The first step in this is to delineate, temporally and spatially, the object of study; i.e. the Metropolitan Greenbelt2. The second is to contextualize historically and disciplinarily those theoretical ideas which conform to the greenbelt concept. The third is to acknowledge, locate and document the elements which have characterized the urban occupation of the Greenbelt and classify these according to the parameters of time, function, urban development and form, which in turn would enable the geography, use, instrumentation and form of its transformation to be analysed on a general metropolitan scale. The fourth step, as a method of verification, is an in-depth analysis of two representative settings within the metropolitan network: the municipality of Pozuelo de Alarcón and the Hortaleza-Barajas district. The content of the document is divided into three parts. Part I focuses on the study’s theoretical foundations, Part II establishes a line of argument at the metropolitan level and Part III examines the phenomenon from a municipal and urban perspective. The thesis, then, begins with a study of the greenbelt concept itself and its meaning, which is far more complex than the accepted instrumental dimension of limiting and containing urbanization in response to the growth of the industrial city of the early 20th century, and which is based on a decentralization of the traditional city. This wider purpose is the setting aside of a large reserved space to safeguard the natural and cultural values of the region and thereby achieve a balance between the city and its residents; that is to say, between man and the space he inhabits. The principal currents of thought in urban planning will then be examined. These have drawn upon a variety of disciplines (economics, sociology, geography, biology, ecology) to develop theories for establishing a new urban order according to the principles of social, economic and environmental equity, and have involved a constant interchange between Europe and the United States. Thus, the City Garden and Regionalist movements would clear the way for proposals such as Greater London and Great Berlin, Chicago and Washington, in which the greenbelt would play a fundamental role. The participation of our own country and the city of Madrid is also discussed, through regional models such as the Besteiro Plan and urban ones like the Bidagor Plan, direct forerunners of 1963’s General Organizational Plan for the Madrid Metropolitan Area, which would set into motion the organization of Madrid’s metropolitan growth. The line of argument followed in this thesis is two-fold: first, an examination of metropolitan development in keeping with the city model as defined in the various General Plans for the development of both the municipality of Madrid and the seven surrounding municipalities which have land designated for its Greenbelt; and second, an examination of this growth in relation to Regional Planning measures, is detailed on a smaller scale (municipal and district), where the conditioning factors affecting the land property structure and the network of biophysical units may be analysed in depth. The first of these is dealt with in Part II and organized into three chapters (4, 5 and 6). Chapter 4 is dedicated to the obligatory starting point of the geographical setting itself. The biophysical characteristics of the territories set aside for the Greenbelt, and which historically have played a role in the area’s exploitation, are described here. These range from expanses of Mediterranean woodland to the north and west of Monte del Pardo to the various types of farmland that have been adapted to the geological substratum and the contours of the terrain (gentle undulations of arkosic sediment in the north, and wide sandy and gypsiferous tableland in the south), as well as orchards planted in low valleys and along watercourses (the creeks of Monte Carmelo, Valdebebas, Quinto, Santo, Butarque, Meaques and Pozuelo). Once this physical reality ha been detailed, in Chapter 5 will examine the various city models proposed by urban planners, both regionally and municipally, in response to the confluence of social, economic and political interests that have characterized the process of occupation in the Greenbelt area during the construction of the MMA. Municipal planning proposals will be collected and examined for the various municipalities which have land designated for the Greenbelt: Madrid, Coslada, Getafe, Leganés, Alcorcón, Boadilla del Monte and Pozuelo de Alarcón. Furthermore, the various territorial organization proposals which have served as references for municipal planning will also be addressed here, in all of their versions –from the sectorial, which have met with more success and institutional approval, to the many attempts at integration, which have been more complex but less influential, precisely for the difficulty of reconciling physical organization with economic development, and private interest with public benefit. The first period in this process was the development of the General Plan of 1963, followed by the first municipal development plans of the 1970s, in which the need for a regional framework that “organized” the territory in an integral fashion was defined. This would serve as a reference for the sectorial actions that marked the metropolitan area’s initial development. The second came two decades later with the approval of the General Plan of 1985, and the network of municipal plans for the surrounding communities, which followed the same philosophy of austerity with regard to territorial growth. The third would begin to take form in 1997, as a new generation of neo-liberal development plans imposed a territorial model based on centralized large-scale metropolitan operations of infrastructure and equipment, which would indiscriminately consume the totality of Madrid’s land. At the end of the Part II, in Chapter 6, the metropolitan area will be represented graphically and the 229 pieces that have been gradually encroaching upon land designated for the Greenbelt will be analysed. This analysis will be carried out according to the parameters defined for the study, and the first general conclusions of the thesis will be based on its findings. It will show how alliances between the various governing authorities in the construction of the city and its environment have successively violated established plans with regard to the definitions of city and territory, how shortages of instrumental and judicial resources have accentuated the dismantling process, and how natural and geographical factors have influenced functional specialization and social segregation in the Madrid Metropolitan Area. The final step, detailed in Part III, will address two fundamental aspects of what has just been described: the lack of consideration for the cultural and environmental values which have shaped this territory and imprinted upon it a specific and unique character; and the structure of land domination, with a precise identification of the social group and agents responsible at each stage of the Greenbelt’s commercialization, who, under an umbrella of social urgency and deceptive public benefit, have used it to obtain substantial financial rewards. For this purpose, a closer look is taken at two specific areas: the municipality of Pozuelo de Alarcón, representative of the suburban growth of an elite population which has occupied the Greenbelt areas of the greatest ecological value; and the Hortaleza-Barajas district, which has offered its territory to large metropolitan business interests, based on activities centred on the connection between Madrid and Barcelona and the system of international air travel. Both of these settings are located to the north of the line which divides the Sierra from La Mancha, and thus occupy the most valuable land in the Madrid region (a functional structure anticipated by Bidagor in 1946, with his city model adapted to the territory of Madrid). Finally, an attempt will be made to interpret the phenomenon of metropolitan Greenbelt occupation, confirming initial hypotheses, specifying the environmental and cultural values that have been violated, and identifying the various players involved, as well as numerous urbanization operations of varying sizes and interests, and the instruments of planning they have used. It will be seen from this that the construction of the MMA has in fact followed a “spread” model, a “grease spot” (as Geddes calls it) which, from the outset of the planning process and according to the definition of a greenbelt as a construction-free zone, was precisely to be avoided. This structural role (to provide a balance between a city and its residents) has thus been abandoned and the Greenbelt converted instead into a communicative framework which, once consolidated, has become the greatest ally of the real estate machine. After this process of violating limits, norms and established behaviour has been described and solidly documented, a reflection will be made on the real influence of the Greenbelt proposal in the construction of the MMA. At the same time, new roles will be suggested for future planning, roles which are deliberate and long term, in opposition to the immediate and circumstantial. This will enable a new interpretation of the theoretical principles behind the greenbelt concept, a space designed to connect the territory of Madrid environmentally, socially and culturally.

Design, Technology and Characterization of Micromechanised Sensors and Actuators for Harsh Environments

Los sistemas micro electro mecánicos (MEMS) han demostrado ser una exitosa familia de dispositivos que pueden usarse como plataforma para el desarrollo de dispositivos con aplicaciones en óptica, comunicaciones, procesado de señal y sensorización. Los dispositivos MEMS estándar suelen estar fabricados usando tecnología de silicio. Sin embargo, el rendimiento de estos MEMS se puede mejorar si se usan otros materiales. Por ejemplo, el diamante nanocristalino (NCD) ofrece unas excelentes propiedades mecánicas, transparencia y una superficie fácil de funcionalizar. Por otro lado, el sistema de materiales (In; Ga; Al)N, los materiales IIIN, se pueden usar para producir estructuras monocristalinas con alta sensibilidad mecánica y química. Además, el AlN se puede depositar por pulverización catódica reactiva sobre varios substratos, incluyendo NCD, para formar capas policristalinas orientadas con alta respuesta piezoeléctrica. Adicionalmente, tanto el NCD como los materiales III-N muestran una gran estabilidad térmica y química, lo que los hace una elección idónea para desarrollar dispositivos para aplicaciones para alta temperatura, ambientes agresivos e incluso para aplicaciones biocompatibles. En esta tesis se han usado estos materiales para el diseño y medición de demostradores tecnológicos. Se han perseguido tres objetivos principales: _ Desarrollo de unos procesos de fabricación apropiados. _ Medición de las propiedades mecánicas de los materiales y de los factores que limitan el rendimiento de los dispositivos. _ Usar los datos medidos para desarrollar dispositivos demostradores complejos. En la primera parte de esta tesis se han estudiado varias técnicas de fabricación. La estabilidad de estos materiales impide el ataque y dificulta la producción de estructuras suspendidas. Los primeros capítulos de esta disertación se dedican al desarrollo de unos procesos de transferencia de patrones por ataque seco y a la optimización del ataque húmedo sacrificial de varios substratos propuestos. Los resultados de los procedimientos de ataque se presentan y se describe la optimización de las técnicas para la fabricación de estructuras suspendidas de NCD y materiales III-N. En un capítulo posterior se estudia el crecimiento de AlN por pulverización catódica. Como se ha calculado en esta disertación para obtener una actuación eficiente de MEMS, las capas de AlN han de ser finas, típicamente d < 200 nm, lo que supone serias dificultades para la obtención de capas orientadas con respuesta piezoeléctrica. Las condiciones de depósito se han mapeado para identificar las fronteras que proporcionan el crecimiento de material orientado desde los primeros pasos del proceso. Además, durante la optimización de los procesos de ataque se estudió un procedimiento para fabricar películas de GaN nanoporoso. Estas capas porosas pueden servir como capas sacrificiales para la fabricación de estructuras suspendidas de GaN con baja tensión residual o como capas para mejorar la funcionalización superficial de sensores químicos o biológicos. El proceso de inducción de poros se discutirá y también se presentarán experimentos de ataque y funcionalización. En segundo lugar, se han determinado las propiedades mecánicas del NCD y de los materiales III-N. Se han fabricado varias estructuras suspendidas para la medición del módulo de Young y de la tensión residual. Además, las estructuras de NCD se midieron en resonancia para calcular el rendimiento de los dispositivos en términos de frecuencia y factor de calidad. Se identificaron los factores intrínsecos y extrínsecos que limitan ambas figuras de mérito y se han desarrollado modelos para considerar estas imperfecciones en las etapas de diseño de los dispositivos. Por otra parte, los materiales III-N normalmente presentan grandes gradientes de deformación residual que causan la deformación de las estructuras al ser liberadas. Se han medido y modelado estos efectos para los tres materiales binarios del sistema para proporcionar puntos de interpolación que permitan predecir las características de las aleaciones del sistema III-N. Por último, los datos recabados se han usado para desarrollar modelos analíticos y numéricos para el diseño de varios dispositivos. Se han estudiado las propiedades de transducción y se proporcionan topologías optimizadas. En el último capítulo de esta disertación se presentan diseños optimizados de los siguientes dispositivos: _ Traviesas y voladizos de AlN=NCD con actuación piezoeléctrica aplicados a nanoconmutadores de RF para señales de alta potencia. _ Membranas circulares de AlN=NCD con actuación piezoeléctrica aplicadas a lentes sintonizables. _ Filtros ópticos Fabry-Pérot basados en cavidades aéreas y membranas de GaN actuadas electrostáticamente. En resumen, se han desarrollado unos nuevos procedimientos optimizados para la fabricación de estructuras de NCD y materiales III-N. Estas técnicas se han usado para producir estructuras que llevaron a la determinación de las principales propiedades mecánicas y de los parámetros de los dispositivos necesarios para el diseño de MEMS. Finalmente, los datos obtenidos se han usado para el diseño optimizado de varios dispositivos demostradores. ABSTRACT Micro Electro Mechanical Systems (MEMS) have proven to be a successful family of devices that can be used as a platform for the development of devices with applications in optics, communications, signal processing and sensorics. Standard MEMS devices are usually fabricated using silicon based materials. However, the performance of these MEMS can be improved if other material systems are used. For instance, nanocrystalline diamond (NCD) offers excellent mechanical properties, optical transparency and ease of surface functionalization. On the other hand, the (In; Ga; Al)N material system, the III-N materials, can be used to produce single crystal structures with high mechanical and chemical sensitivity. Also, AlN can be deposited by reactive sputtering on various substrates, including NCD, to form oriented polycrystalline layers with high piezoelectric response. In addition, both NCD and III-N materials exhibit high thermal and chemical stability, which makes these material the perfect choice for the development of devices for high temperatures, harsh environments and even biocompatible applications. In this thesis these materials have been used for the design and measurement of technological demonstrators. Three main objectives have been pursued: _ Development of suitable fabrication processes. _ Measurement of the material mechanical properties and device performance limiting factors. _ Use the gathered data to design complex demonstrator devices. In a first part of the thesis several fabrication processes have been addressed. The stability of these materials hinders the etching of the layers and hampers the production of free standing structures. The first chapters of this dissertation are devoted to the development of a dry patterning etching process and to sacrificial etching optimization of several proposed substrates. The results of the etching processes are presented and the optimization of the technique for the manufacturing of NCD and III-N free standing structures is described. In a later chapter, sputtering growth of thin AlN layers is studied. As calculated in this dissertation, for efficient MEMS piezoelectric actuation the AlN layers have to be very thin, typically d < 200 nm, which poses serious difficulties to the production of c-axis oriented material with piezoelectric response. The deposition conditions have been mapped in order to identify the boundaries that give rise to the growth of c-axis oriented material from the first deposition stages. Additionally, during the etching optimization a procedure for fabricating nanoporous GaN layers was also studied. Such porous layers can serve as a sacrificial layer for the release of low stressed GaN devices or as a functionalization enhancement layer for chemical and biological sensors. The pore induction process will be discussed and etching and functionalization trials are presented. Secondly, the mechanical properties of NCD and III-N materials have been determined. Several free standing structures were fabricated for the measurement of the material Young’s modulus and residual stress. In addition, NCD structures were measured under resonance in order to calculate the device performance in terms of frequency and quality factor. Intrinsic and extrinsic limiting factors for both figures were identified and models have been developed in order to take into account these imperfections in the device design stages. On the other hand, III-N materials usually present large strain gradients that lead to device deformation after release. These effects have been measured and modeled for the three binary materials of the system in order to provide the interpolation points for predicting the behavior of the III-N alloys. Finally, the gathered data has been used for developing analytic and numeric models for the design of various devices. The transduction properties are studied and optimized topologies are provided. Optimized design of the following devices is presented at the last chapter of this dissertation: _ AlN=NCD piezoelectrically actuated beams applied to RF nanoswitches for large power signals. _ AlN=NCD piezoelectrically actuated circular membranes applied to tunable lenses. _ GaN based air gap tunable optical Fabry-Pérot filters with electrostatic actuation. On the whole, new optimized fabrication processes has been developed for the fabrication of NCD and III-N MEMS structures. These processing techniques was used to produce structures that led to the determination of the main mechanical properties and device parameters needed for MEMS design. Lastly, the gathered data was used for the design of various optimized demonstrator devices.

Demonstration of an interferon γ-dependent tumor surveillance system in immunocompetent mice

This study demonstrates that endogenously produced interferon γ (IFN-γ) forms the basis of a tumor surveillance system that controls development of both chemically induced and spontaneously arising tumors in mice. Compared with wild-type mice, mice lacking sensitivity to either IFN-γ (i.e., IFN-γ receptor-deficient mice) or all IFN family members (i.e., Stat1-deficient mice) developed tumors more rapidly and with greater frequency when challenged with different doses of the chemical carcinogen methylcholanthrene. In addition, IFN-γ-insensitive mice developed tumors more rapidly than wild-type mice when bred onto a background deficient in the p53 tumor-suppressor gene. IFN-γ-insensitive p53−/− mice also developed a broader spectrum of tumors compared with mice lacking p53 alone. Using tumor cells derived from methylcholanthrene-treated IFN-γ-insensitive mice, we found IFN-γ’s actions to be mediated at least partly through its direct effects on the tumor cell leading to enhanced tumor cell immunogenicity. The importance and generality of this system is evidenced by the finding that certain types of human tumors become selectively unresponsive to IFN-γ. Thus, IFN-γ forms the basis of an extrinsic tumor-suppressor mechanism in immunocompetent hosts.

Chromophore structural changes in rhodopsin from nanoseconds to microseconds following pigment photolysis

Rhodopsin is a prototypical G protein-coupled receptor that is activated by photoisomerization of its 11-cis-retinal chromophore. Mutant forms of rhodopsin were prepared in which the carboxylic acid counterion was moved relative to the positively charged chromophore Schiff base. Nanosecond time-resolved laser photolysis measurements of wild-type recombinant rhodopsin and two mutant pigments then were used to determine reaction schemes and spectra of their early photolysis intermediates. These results, together with linear dichroism data, yielded detailed structural information concerning chromophore movements during the first microsecond after photolysis. These chromophore structural changes provide a basis for understanding the relative movement of rhodopsin’s transmembrane helices 3 and 6 required for activation of rhodopsin. Thus, early structural changes following isomerization of retinal are linked to the activation of this G protein-coupled receptor. Such rapid structural changes lie at the heart of the pharmacologically important signal transduction mechanisms in a large variety of receptors, which use extrinsic activators, but are impossible to study in receptors using diffusible agonist ligands.

Response-reinforcement learning is dependent on N-methyl-d-aspartate receptor activation in the nucleus accumbens core

The nucleus accumbens, a site within the ventral striatum, is best known for its prominent role in mediating the reinforcing effects of drugs of abuse such as cocaine, alcohol, and nicotine. Indeed, it is generally believed that this structure subserves motivated behaviors, such as feeding, drinking, sexual behavior, and exploratory locomotion, which are elicited by natural rewards or incentive stimuli. A basic rule of positive reinforcement is that motor responses will increase in magnitude and vigor if followed by a rewarding event. It is likely, therefore, that the nucleus accumbens may serve as a substrate for reinforcement learning. However, there is surprisingly little information concerning the neural mechanisms by which appetitive responses are learned. In the present study, we report that treatment of the nucleus accumbens core with the selective competitive N-methyl-d-aspartate (NMDA) antagonist 2-amino-5-phosphonopentanoic acid (AP-5; 5 nmol/0.5 μl bilaterally) impairs response-reinforcement learning in the acquisition of a simple lever-press task to obtain food. Once the rats learned the task, AP-5 had no effect, demonstrating the requirement of NMDA receptor-dependent plasticity in the early stages of learning. Infusion of AP-5 into the accumbens shell produced a much smaller impairment of learning. Additional experiments showed that AP-5 core-treated rats had normal feeding and locomotor responses and were capable of acquiring stimulus-reward associations. We hypothesize that stimulation of NMDA receptors within the accumbens core is a key process through which motor responses become established in response to reinforcing stimuli. Further, this mechanism, may also play a critical role in the motivational and addictive properties of drugs of abuse.

Photoassembly of the manganese cluster and oxygen evolution from monomeric and dimeric CP47 reaction center photosystem II complexes

Isolated subcomplexes of photosystem II from spinach (CP47RC), composed of D1, D2, cytochrome b559, CP47, and a number of hydrophobic small subunits but devoid of CP43 and the extrinsic proteins of the oxygen-evolving complex, were shown to reconstitute the Mn4Ca1Clx cluster of the water-splitting system and to evolve oxygen. The photoactivation process in CP47RC dimers proceeds by the same two-step mechanism as observed in PSII membranes and exhibits the same stoichiometry for Mn2+, but with a 10-fold lower affinity for Ca2+ and an increased susceptibility to photodamage. After the lower Ca2+ affinity and the 10-fold smaller absorption cross-section for photons in CP47 dimers is taken into account, the intrinsic rate constant for the rate-limiting calcium-dependent dark step is indistinguishable for the two systems. The monomeric form of CP47RC also showed capacity to photoactivate and catalyze water oxidation, but with lower activity than the dimeric form and increased susceptibility to photodamage. After optimization of the various parameters affecting the photoactivation process in dimeric CP47RC subcores, 18% of the complexes were functionally reconstituted and the quantum efficiency for oxygen production by reactivated centers approached 96% of that observed for reconstituted photosystem II-enriched membranes.

Rad18 Is Required for DNA Repair and Checkpoint Responses in Fission Yeast

To survive damage to the genome, cells must respond by activating both DNA repair and checkpoint responses. Using genetic screens in the fission yeast Schizosaccharomyces pombe, we recently isolated new genes required for DNA damage checkpoint control. We show here that one of these strains defines a new allele of the previously described rad18 gene, rad18-74. rad18 is an essential gene, even in the absence of extrinsic DNA damage. It encodes a conserved protein related to the structural maintenance of chromosomes proteins. Point mutations in rad18 lead to defective DNA repair pathways responding to both UV-induced lesions and, as we show here, double-stranded breaks. Furthermore, rad18p is required to maintain cell cycle arrest in the presence of DNA damage, and failure of this leads to highly aberrant mitoses. A gene encoding a BRCT-containing protein, brc1, was isolated as an allele-specific high-copy suppressor of rad18-74. brc1 is required for mitotic fidelity and for cellular viability in strains with rad18 mutations but is not essential for DNA damage responses. Mutations in rad18 and brc1 are synthetically lethal with a topoisomerase II mutant (top2-191), indicating that these proteins play a role in chromatin organization. These studies show a role for chromatin organization in the maintenance or activation of responses to DNA damage.

Region-specific differentiation of neural tube-derived neuronal restricted progenitor cells after heterotopic transplantation

Spinal cord neuronal restricted progenitor (NRP) cells, when transplanted into the neonatal anterior forebrain subventricular zone, migrate to distinct regions throughout the forebrain including the olfactory bulb, frontal cortex, and occipital cortex but not to the hippocampus. Their migration pattern and differentiation potential is distinct from anterior forebrain subventricular zone NRPs. Irrespective of their final destination, NRP cells do not differentiate into glia. Rather they synthesize neurotransmitters, acquire region-specific phenotypes, and receive synapses from host neurons after transplantation. Spinal cord NRPs express choline acetyl transferase even in regions where host neurons do not express this marker. The restricted distribution of transplanted spinal cord NRP cells and their acquisition of varied region-specific phenotypes suggest that their ultimate fate and phenotype is dictated by a combination of intrinsic properties and extrinsic cues from the host.

Redox state is a central modulator of the balance between self-renewal and differentiation in a dividing glial precursor cell

We have discovered that intracellular redox state appears to be a necessary and sufficient modulator of the balance between self-renewal and differentiation in dividing oligodendrocyte-type-2 astrocyte progenitor cells. The intracellular redox state of freshly isolated progenitors allows prospective isolation of cells with different self-renewal characteristics. Redox state is itself modulated by cell-extrinsic signaling molecules that alter the balance between self-renewal and differentiation: growth factors that promote self-renewal cause progenitors to become more reduced, while signaling molecules that promote differentiation cause progenitors to become more oxidized. Moreover, pharmacological antagonists of the redox effects of these cell-extrinsic signaling molecules antagonize their effects on self-renewal and differentiation, indicating that cell-extrinsic signaling molecules that modulate this balance converge on redox modulation as a critical component of their effector mechanism.

Multiple benefits of multiple mating in guppies

The rewards of promiscuity for males are undisputed. But why should a female mate promiscuously, particularly when her partners offer no resources other than sperm and increase her chances of succumbing to predation or disease? This question has been hotly debated but at present remains largely unresolved [Jennions, M. D. & Petrie, M. (2000) Biol. Rev. 75, 21–64]. One possibility is that females exploit postcopulatory mechanisms, such as sperm competition, to increase both the quality and quantity of their offspring. In this paper, we use the Trinidadian guppy, a species with a resource-free mating system, to test the hypothesis that females gain multiple benefits from multiple mating. Our results indicate that multiply mated females secure substantive advantages: They have shorter gestation times and larger broods, and they produce offspring with better developed schooling abilities and escape responses than their singly mated counterparts.

Oocytes are a source of catecholamines in the primate ovary: Evidence for a cell–cell regulatory loop

Catecholamines, thought to derive from the extrinsic innervation of the ovary, participate in the regulation of ovarian development and mature gonadal function. Recently, intraovarian neurons containing tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine biosynthesis, were described in the ovary of nonhuman primates. We now show that the primate ovary expresses both the genes encoding TH and dopamine β-hydroxylase (DBH), the key enzymes in norepinephrine (NE) biosynthesis. Ovarian neurons were identified as a site of TH and DBH gene expression, and surprisingly, oocytes were identified as an exclusive site of DBH synthesis. Oocytes contain neither TH mRNA nor protein, indicating that they are unable to synthesize dopamine (DA). They did, however, express a DA transporter gene identical to that found in human brain. The physiological relevance of this transporter system and DBH in oocytes was indicated by the ability of isolated oocytes to metabolize exogenous DA into NE. Isolated follicles containing oocytes—but not those from which the oocytes had been removed—responded to DA with an elevation in cAMP levels; this elevation was prevented by propranolol, a β-adrenoreceptor antagonist. The results suggest that oocytes and somatic cells are linked by a neuroendocrine loop consisting of NE synthesized in oocytes from actively transported DA and cAMP produced by somatic follicular cells in response to NE-induced β-adrenoreceptor activation.

Arabidopsis Sec21p and Sec23p Homologs. Probable Coat Proteins of Plant COP-Coated Vesicles1

Intracellular protein transport between the endoplasmic reticulum (ER) and the Golgi apparatus and within the Golgi apparatus is facilitated by COP (coat protein)-coated vesicles. Their existence in plant cells has not yet been demonstrated, although the GTP-binding proteins required for coat formation have been identified. We have generated antisera against glutathione-S-transferase-fusion proteins prepared with cDNAs encoding the Arabidopsis Sec21p and Sec23p homologs (AtSec21p and AtSec23p, respectively). The former is a constituent of the COPI vesicle coatomer, and the latter is part of the Sec23/24p dimeric complex of the COPII vesicle coat. Cauliflower (Brassica oleracea) inflorescence homogenates were probed with these antibodies and demonstrated the presence of AtSec21p and AtSec23p antigens in both the cytosol and membrane fractions of the cell. The membrane-associated forms of both antigens can be solubilized by treatments typical for extrinsic proteins. The amounts of the cytosolic antigens relative to the membrane-bound forms increase after cold treatment, and the two antigens belong to different protein complexes with molecular sizes comparable to the corresponding nonplant coat proteins. Sucrose-density-gradient centrifugation of microsomal cell membranes from cauliflower suggests that, although AtSec23p seems to be preferentially associated with ER membranes, AtSec21p appears to be bound to both the ER and the Golgi membranes. This could be in agreement with the notion that COPII vesicles are formed at the ER, whereas COPI vesicles can be made by both Golgi and ER membranes. Both AtSec21p and AtSec23p antigens were detected on membranes equilibrating at sucrose densities equivalent to those typical for in vitro-induced COP vesicles from animal and yeast systems. Therefore, a further purification of the putative plant COP vesicles was undertaken.

Growth factors regulate phototransduction in retinal rods by modulating cyclic nucleotide-gated channels through dephosphorylation of a specific tyrosine residue

Illumination of vertebrate rod photoreceptors leads to a decrease in the cytoplasmic cGMP concentration and closure of cyclic nucleotide-gated (CNG) channels. Except for Ca2+, which plays a negative feedback role in adaptation, and 11-cis-retinal, supplied by the retinal pigment epithelium, all of the biochemical machinery of phototransduction is thought to be contained within rod outer segments without involvement of extrinsic regulatory molecules. Here we show that insulin-like growth factor-I (IGF-I), a paracrine factor released from the retinal pigment epithelium, alters phototransduction by rapidly increasing the cGMP sensitivity of CNG channels. The IGF-I-signaling pathway ultimately involves a protein tyrosine phosphatase that catalyzes dephosphorylation of a specific residue in the α-subunit of the rod CNG channel protein. IGF-I conjointly accelerates the kinetics and increases the amplitude of the light response, distinct from events that accompany adaptation. These effects of IGF-I could result from the enhancement of the cGMP sensitivity of CNG channels. Hence, in addition to long-term control of development and survival of rods, growth factors regulate phototransduction in the short term by modulating CNG channels.

Kinetic proofreading models for cell signaling predict ways to escape kinetic proofreading

In the context of cell signaling, kinetic proofreading was introduced to explain how cells can discriminate among ligands based on a kinetic parameter, the ligand-receptor dissociation rate constant. In the kinetic proofreading model of cell signaling, responses occur only when a bound receptor undergoes a complete series of modifications. If the ligand dissociates prematurely, the receptor returns to its basal state and signaling is frustrated. We extend the model to deal with systems where aggregation of receptors is essential to signal transduction, and present a version of the model for systems where signaling depends on an extrinsic kinase. We also investigate the kinetics of signaling molecules, “messengers,” that are generated by aggregated receptors but do not remain associated with the receptor complex. We show that the extended model predicts modes of signaling that exhibit kinetic discrimination for some range of parameters but for other parameter values show little or no discrimination and thus escape kinetic proofreading. We compare model predictions with experimental data.