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.

High-Throughput Sequencing of microRNAs in Peripheral Blood Mononuclear Cells: Identification of Potential Weight Loss Biomarkers

INTRODUCTION: MicroRNAs (miRNAs) are being increasingly studied in relation to energy metabolism and body composition homeostasis. Indeed, the quantitative analysis of miRNAs expression in different adiposity conditions may contribute to understand the intimate mechanisms participating in body weight control and to find new biomarkers with diagnostic or prognostic value in obesity management. OBJECTIVE: The aim of this study was the search for miRNAs in blood cells whose expression could be used as prognostic biomarkers of weight loss. METHODS: Ten Caucasian obese women were selected among the participants in a weight-loss trial that consisted in following an energy-restricted treatment. Weight loss was considered unsuccessful when <5% of initial body weight (non-responders) and successful when >5% (responders). At baseline, total miRNA isolated from peripheral blood mononuclear cells (PBMC) was sequenced with SOLiD v4. The miRNA sequencing data were validated by RT-PCR. RESULTS: Differential baseline expression of several miRNAs was found between responders and non-responders. Two miRNAs were up-regulated in the non-responder group (mir-935 and mir-4772) and three others were down-regulated (mir-223, mir-224 and mir-376b). Both mir-935 and mir-4772 showed relevant associations with the magnitude of weight loss, although the expression of other transcripts (mir-874, mir-199b, mir-766, mir-589 and mir-148b) also correlated with weight loss. CONCLUSIONS: This research addresses the use of high-throughput sequencing technologies in the search for miRNA expression biomarkers in obesity, by determining the miRNA transcriptome of PBMC. Basal expression of different miRNAs, particularly mir-935 and mir-4772, could be prognostic biomarkers and may forecast the response to a hypocaloric diet.

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.

SNP Discovery in European Anchovy (Engraulis encrasicolus, L) by High-Throughput Transcriptome and Genome Sequencing

12 p.

Biocompatible Bacterial Cellulose-Poly(2-hydroxyethyl methacrylate) Nanocomposite Films

A series of bacterial cellulose-poly(2-hydroxyethyl methacrylate) nanocomposite films was prepared by in situ radical polymerization of 2-hydroxyethyl methacrylate (HEMA), using variable amounts of poly(ethylene glycol) diacrylate (PEGDA) as crosslinker. Thin films were obtained, and their physical, chemical, thermal, and mechanical properties were evaluated. The films showed improved translucency compared to BC and enhanced thermal stability and mechanical performance when compared to poly(2-hydroxyethyl methacrylate) (PHEMA). Finally, BC/PHEMA nanocomposites proved to be nontoxic to human adipose-derived mesenchymal stem cells (ADSCs) and thus are pointed as potential dry dressings for biomedical applications.

Adenylate Cyclase Toxin promotes bacterial internalisation into non phagocytic cells

Bordetella pertussis causes whooping cough, a respiratory infectious disease that is the fifth largest cause of vaccine-preventable death in infants. Though historically considered an extracellular pathogen, this bacterium has been detected both in vitro and in vivo inside phagocytic and non-phagocytic cells. However the precise mechanism used by B. pertussis for cell entry, or the putative bacterial factors involved, are not fully elucidated. Here we find that adenylate cyclase toxin (ACT), one of the important toxins of B. pertussis, is sufficient to promote bacterial internalisation into non-phagocytic cells. After characterization of the entry route we show that uptake of "toxin-coated bacteria" proceeds via a clathrin-independent, caveolae-dependent entry pathway, allowing the internalised bacteria to survive within the cells. Intracellular bacteria were found inside non-acidic endosomes with high sphingomyelin and cholesterol content, or "free" in the cytosol of the invaded cells, suggesting that the ACT-induced bacterial uptake may not proceed through formation of late endolysosomes. Activation of Tyr kinases and toxin-induced Ca2+-influx are essential for the entry process. We hypothesize that B. pertussis might use ACT to activate the endocytic machinery of non-phagocytic cells and gain entry into these cells, in this way evading the host immune system.

Amplicon metagenomics for elucidating microbial community changes associated to conventional and organic farming of Beta vulgaris and Solanum lycopersicum

Soil microbial community changes associated to conventional and organic farming of two relevant crops (Beta vulgaris and Solanum lycopersicum) were analysed through 16s rRNA amplicon sequencing. This study revealed microbial communities in the agricultural soils studied to be similar to other reported nutrient-rich microbiomes, and some significant differences between the microbial communities associated to the two farming practices were found. Some phyla (Chloroflexi and Thermi) were found to be present in different abundances according to soil treatment. As chloroplast interference can be a stumbling block in plant-associated 16s rRNA amplicon metagenomics analysis of aerial plant tissues, two protocols for bacterial cell detachment (orbital shaking and ultrasound treatment) and two protocols for microbial biomass recovery (centrifugation and filtration) were tested regarding their efficiency at excluding plant-DNA. An alternative method to the one proposed by Rastogi et al (2010) for evaluating the chloroplast-amplicon content in post-PCR samples was tested, and the method revealed that filtration was the most efficient protocol in minimising chloroplast interference.

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.