Dialog Systems based on Markov decision processes over two real tasks

In this work the state of the art of the automatic dialogue strategy management using Markov decision processes (MDP) with reinforcement learning (RL) is described. Partially observable Markov decision processes (POMDP) are also described. To test the validity of these methods, two spoken dialogue systems have been developed. The first one is a spoken dialogue system for weather forecast providing, and the second one is a more complex system for train information. With the first system, comparisons between a rule-based system and an automatically trained system have been done, using a real corpus to train the automatic strategy. In the second system, the scalability of these methods when used in larger systems has been tested.

Testu arteko koherentziazko erlazio-egitura: lehen urratsak euskaraz

[EU]Testu bat koherente egiten duten arrazoiak ulertzea oso baliagarria da testuaren beraren ulermenerako, koherentzia eta koherentzia-erlazioak testu bat edo gehiago koherente diren ondorioztatzen laguntzen baitigu. Lan honetan gai bera duten testu ezberdinen arteko koherentziazko 3 Cross Document Structure Theory edo CST (Radev, 2000) erlazio aztertu eta sailkatu dira. Hori egin ahal izateko, euskaraz idatziriko gai berari buruzko testuak segmentatzeko eta beraien arteko erlazioak etiketatzeko gidalerroak proposatzen dira. 10 testuz osaturiko corpusa etiketatu da; horietako 3 cluster bi etiketatzailek aztertu dute. Etiketatzaileen arteko adostasunaren berri ematen dugu. Koherentzia-erlazioak garatzea oso garrantzitsua da Hizkuntzaren Prozesamenduko hainbat sistementzat, hala nola, informazioa erauzteko sistementzat, itzulpen automatikoarentzat, galde-erantzun sistementzat eta laburpen automatikoarentzat. Etorkizunean CSTko erlazio guztiak corpus esanguratsuan aztertuko balira, testuen arteko koherentzia- erlazioak euskarazko testuen prozesaketa automatikoa bideratzeko lehenengo pausua litzateke hemen egindakoa.

Feature Selection for Speech Emotion Recognition in Spanish and Basque: On the Use of Machine Learning to Improve Human-Computer Interaction

Study of emotions in human-computer interaction is a growing research area. This paper shows an attempt to select the most significant features for emotion recognition in spoken Basque and Spanish Languages using different methods for feature selection. RekEmozio database was used as the experimental data set. Several Machine Learning paradigms were used for the emotion classification task. Experiments were executed in three phases, using different sets of features as classification variables in each phase. Moreover, feature subset selection was applied at each phase in order to seek for the most relevant feature subset. The three phases approach was selected to check the validity of the proposed approach. Achieved results show that an instance-based learning algorithm using feature subset selection techniques based on evolutionary algorithms is the best Machine Learning paradigm in automatic emotion recognition, with all different feature sets, obtaining a mean of 80,05% emotion recognition rate in Basque and a 74,82% in Spanish. In order to check the goodness of the proposed process, a greedy searching approach (FSS-Forward) has been applied and a comparison between them is provided. Based on achieved results, a set of most relevant non-speaker dependent features is proposed for both languages and new perspectives are suggested.

EUSMT: incorporating linguistic information to SMT for a morphologically rich language. Its use in SMT-RBMT-EBMT hybridation

148 p.: graf.

Candida albicans Increases Tumor Cell Adhesion to Endothelial Cells In Vitro: Intraspecific Differences and Importance of the Mannose Receptor

The dimorphic fungus Candida albicans is able to trigger a cytokine-mediated pro-inflammatory response that increases tumor cell adhesion to hepatic endothelium and metastasis. To check the intraspecific differences in this effect, we used an in vitro murine model of hepatic response against C. albicans, which made clear that tumor cells adhered more to endothelium incubated with blastoconidia, both live and killed, than germ tubes. This finding was related to the higher carbohydrate/protein ratio found in blastoconidia. In fact, destruction of mannose ligand residues on the cell surface by metaperiodate treatment significantly reduced tumor cell adhesion induced. Moreover, we also noticed that the effect of clinical strains was greater than that of the reference one. This finding could not be explained by the carbohydrate/protein data, but to explain these differences between strains, we analyzed the expression level of ten genes (ADH1, APE3, IDH2, ENO1, FBA1, ILV5, PDI1, PGK1, QCR2 and TUF1) that code for the proteins identified previously in a mannoprotein-enriched pro-metastatic fraction of C. albicans. The results corroborated that their expression was higher in clinical strains than the reference one. To confirm the importance of the mannoprotein fraction, we also demonstrate that blocking the mannose receptor decreases the effect of C. albicans and its mannoproteins, inhibiting IL-18 synthesis and tumor cell adhesion increase by around 60%. These findings could be the first step towards a new treatment for solid organ cancers based on the role of the mannose receptor in C. albicans-induced tumor progression and metastasis.

Semi-automatic generation of learning domain modules for technology supported learning systems

In a time when Technology Supported Learning Systems are being widely used, there is a lack of tools that allows their development in an automatic or semi-automatic way. Technology Supported Learning Systems require an appropriate Domain Module, ie. the pedagogical representation of the domain to be mastered, in order to be effective. However, content authoring is a time and effort consuming task, therefore, efforts in automatising the Domain Module acquisition are necessary.Traditionally, textbooks have been used as the main mechanism to maintain and transmit the knowledge of a certain subject or domain. Textbooks have been authored by domain experts who have organised the contents in a means that facilitate understanding and learning, considering pedagogical issues.Given that textbooks are appropriate sources of information, they can be used to facilitate the development of the Domain Module allowing the identification of the topics to be mastered and the pedagogical relationships among them, as well as the extraction of Learning Objects, ie. meaningful fragments of the textbook with educational purpose.Consequently, in this work DOM-Sortze, a framework for the semi-automatic construction of Domain Modules from electronic textbooks, has been developed. DOM-Sortze uses NLP techniques, heuristic reasoning and ontologies to fulfill its work. DOM-Sortze has been designed and developed with the aim of automatising the development of the Domain Module, regardless of the subject, promoting the knowledge reuse and facilitating the collaboration of the users during the process.

On the Selection of Non-Invasive Methods Based on Speech Analysis Oriented to Automatic Alzheimer Disease Diagnosis

The work presented here is part of a larger study to identify novel technologies and biomarkers for early Alzheimer disease (AD) detection and it focuses on evaluating the suitability of a new approach for early AD diagnosis by non-invasive methods. The purpose is to examine in a pilot study the potential of applying intelligent algorithms to speech features obtained from suspected patients in order to contribute to the improvement of diagnosis of AD and its degree of severity. In this sense, Artificial Neural Networks (ANN) have been used for the automatic classification of the two classes (AD and control subjects). Two human issues have been analyzed for feature selection: Spontaneous Speech and Emotional Response. Not only linear features but also non-linear ones, such as Fractal Dimension, have been explored. The approach is non invasive, low cost and without any side effects. Obtained experimental results were very satisfactory and promising for early diagnosis and classification of AD patients.

Weighting social preferences in participatory multi-criteria evaluations: a case study on sustainable natural resource management

28 p.

Gender segregation in technological undergraduate studies in the University of the Basque Country (UPV/EHU)

In this work we study the gender segregation in technological undergraduate studies in the University of The Basque Country (UPV/EHU). For this study we use the data of new admissions at the UPV/EHU. They are from the time period of the years 2003-2013. We focus on the first and last year to check if the segregation has changed over these ten years. We build segregation curves within the Lorenz approach. Our results show that the gender segregation in technological undergraduate studies in the University of the Basque Country has increased over the last ten years. We also show that the distribution between men and women has changed.  

Gender equality in life insurance

[EN]This project is going to study the implications of the gender of an individual in the rate-setting process of life insurance. In order to do so there is a review of the continuous changes that have taken place in the national and European legislation following the enactment of the Directive 2004/113/EC, as well as its consequences from the prohibition to differentiate the premiums and benefits on the grounds of gender. In this area, the evolution of the Spanish insurance sector and the influence of the new legislation are examined. Furthermore, there is an analysis of the differences between men and women, which to some extent have a direct impact in the management and development of the life insurance companies. Finally, methods to calculate the premium and the benefits are proposed with the purpose of preventing the restrictions imposed by the Directive 2004/113/EC. In order to check the repercussions of the use of unisex tables a comparison is made between the premiums obtained for a whole life insurance by allocating the same weighing to the actuarial male and female mortality tables and those that would result if the distinction by gender were allowed.

Increased antiparkinson efficacy of the combined administration of VEGF- and GDNF-loaded nanospheres in a partial lesion model of Parkinson's disease

Current research efforts are focused on the application of growth factors, such as glial cell line-derived neurotrophic factor (GDNF) and vascular endothelial growth factor (VEGF), as neuroregenerative approaches that will prevent the neurodegenerative process in Parkinson's disease. Continuing a previous work published by our research group, and with the aim to overcome different limitations related to growth factor administration, VEGF and GDNF were encapsulated in poly(lactic-co-glycolic acid) nanospheres (NS). This strategy facilitates the combined administration of the VEGF and GDNF into the brain of 6-hydroxydopamine (6-OHDA) partially lesioned rats, resulting in a continuous and simultaneous drug release. The NS particle size was about 200 nm and the simultaneous addition of VEGF NS and GDNF NS resulted in significant protection of the PC-12 cell line against 6-OHDA in vitro. Once the poly(lactic-co-glycolic acid) NS were implanted into the striatum of 6-OHDA partially lesioned rats, the amphetamine rotation behavior test was carried out over 10 weeks, in order to check for in vivo efficacy. The results showed that VEGF NS and GDNF NS significantly decreased the number of amphetamine-induced rotations at the end of the study. In addition, tyrosine hydroxylase immunohistochemical analysis in the striatum and the external substantia nigra confirmed a significant enhancement of neurons in the VEGF NS and GDNF NS treatment group. The synergistic effect of VEGF NS and GDNF NS allows for a reduction of the dose by half, and may be a valuable neurogenerative/neuroreparative approach for treating Parkinson's disease.

Energy retrofits in social housing. Analysis of its thermal behaviour.

468 p.