Estudio bibliográfico de sistemas educativos en el ámbito de la inteligencia artificial

Este informe trata de reunir en un documento la revisión bibliográfica realizada como base necesaria para el desarrollo de una tesis doctoral sobre las interacciones entre profesores y alumnos en las sesiones tradicionales de aprendizaje. Nuestra investigación se enmarca en los trabajos del grupo GALAN de la Facultad de Informática de San Sebastián de la UPV-EHU, que se dedica desde sus comienzos al desarrollo de herramientas educativas flexibles dotadas de comportamiento inteligente. En este informe presentamos un estudio bibliográfico de sistemas educativos en el ámbito de la inteligencia artificial. En particular se centra en los siguientes aspectos: tutores inteligentes, sistemas educativos para la web y herramientas de ayuda al profesor. Además se incluye una revisión del modelado de usuario y un estudio de técnicas para el análisis de datos.

Influence of Training Set Selection in Artificial Neural Network-Based Propagation Path Loss Predictions

This paper analyzes the use of artificial neural networks (ANNs) for predicting the received power/path loss in both outdoor and indoor links. The approach followed has been a combined use of ANNs and ray-tracing, the latter allowing the identification and parameterization of the so-called dominant path. A complete description of the process for creating and training an ANN-based model is presented with special emphasis on the training process. More specifically, we will be discussing various techniques to arrive at valid predictions focusing on an optimum selection of the training set. A quantitative analysis based on results from two narrowband measurement campaigns, one outdoors and the other indoors, is also presented.

Técnicas de Inteligencia Artificial integradas en una aplicación de gestión para una casa rural en JQueryMobile y Protégé Frames.

Este proyecto plantea el reto de realizar una aplicación web para la gestión y control de una casa rural desde dispositivos móviles como smartphones y tabletas, con una interfaz de propietario para gestionar de manera dinámica las diferentes partes de la misma, así como la inclusión de una zona de niños donde se aplicarán tecnologías de Inteligencia Artificial en concreto representación del conocimiento mediante “frames”, donde se permitirá a los usuarios realizar preguntas al sistema para intentar adivinar un árbol que previamente la propietaria de la casa rural habrá establecido. El objetivo no es elaborar un sistema experto, tarea que requeriría muchas más horas que las que corresponden a un proyecto de fin de carrera, sino comprobar la posibilidad de integración de estas herramientas en una aplicación orientada a dispositivos móviles. Se emplearán las funcionalidades de HTML5 para la inclusión de la “zona explorador” donde los niños podrán geo localizar árboles, así como su posterior búsqueda a modo de GPS donde los usuarios podrán ver donde está situado el árbol que desean buscar y su posición actual, la cual se irá actualizando automáticamente.

SYSTEMATICS OF GROUNDWATER OLIGOCHAETES (ANNELIDA, CLITELLATA) IN KARSTIC AREAS OF THE CANTABRIAN REGION/ ERREGIO KANTAURIARREKO EREMU KARSTIKOETAN DAUDEN LURPEKO OLIGOKETO (ANNELIDA, CLITELLATA) URT

This thesis deals with the oligochaete taxa (Annelida, Clitellata) from several karst units in the Cantabrian region, northern Iberian Peninsula. Groundwater oligochaetes are still poorly known fauna and the area seems to be a promising hotspot for groundwater taxa. Metodology is based on both morphological and molecular analyses. More than 7,000 specimens were collected from five karst units and >60 taxa were identified. Stygobiont oligochaete fauna in the northern Iberian Peninsula is diverse and mostly endemic (range areas <300 km). Three new stygogiont oligochaete species are described: Gianius navarroi Rodriguez & Achurra, 2010, Isochaetides gianii Rodriguez & Achurra, 2010 and Troglodrilus jugeti Achurra et al., 2012; and another four new taxa will be described in the near future. Taxonomic remarks on Lophochaeta ignota Stolc, 1886 and Troglodrilus galarzai (Giani & Rodriguez, 1988) are provided. The controversial separation of L. ignota and Heterochaeta costata Claparède, 1863 from Tubifex Lamarck, 1816 is corroborated by mitochondrial molecular data. Following the DNA barcoding method, individuals of the stygoxene species Stylodrilus heringianus Claparède, 1862 from different geographic areas are shown to represent a single metapopulation. The first phylogenetic analysis of the subfamily Tubificinae based on molecular data is attemped, which although incompletely resolved, evidences for the first time a close relationship between a stygobiont oligochaete (Troglodrilus Juget et al. 2006) and an estuarine especies (Heterochaeta costata Claparéde, 1863). A marine ancestor is hypothesised for Troglodrilus. Finally, Ereñozar karst unit (Biscay) is suggested to be a hotspot for groundwater oligochaetes (11 stygobiont taxa, of which 4 are endemic to the karst unit) and several biodiversity indices (Species richness, Rarity, Vulnerability and Complementarity) are shown to be useful tools for conservation management of groundwater habitats in that karst area.

Building up a frequency stabilisation based on a wavemeter for a diode laser and characterisation for a digital mirror device for artificial potentials

It has built and characterised a laser and It has learned what each of the components does. It has been able to run the laser in single-mode and stabilised it around a desired setpoint thanks to a PID controller that It has programmed. It has established a communication between the PID controller programmed in LabVIEW and Arduino Due, the DAC that It has chosen after comparing it with another candidate. It has learned some basics of how the LightCrafter 4500 DMD works. The projected light is the composition of the lights of three LED’s, each of which has a certain on-time. The mirrors chose to be in on- or off-stages depending to the amount of intensity that we want for each colour.