Teletandem and performativity

Teletandem is a virtual autonomous VoIP2 technology-based context (webcam images, voice and text). Within this context, two students help each other learn their native (or other) language through intercultural and linguistic collaboration. Performative Theory can shed light on the constitution of these students' national identities, as they engage in linguistic performances of marking and discussing differences between their countries during teletandem. Based on critical approaches to discourse and intercultural communication, my analysis shows that this online intercultural contact opens innovative possibilities for foreign language teachers to promote intercultural contact with the different. However, without teacher mediation, teletandem interactions may fall into shallow performances of sedimented and pre-given representations of self and other. Subsequently, this article concludes with a discussion of relevant pedagogical points for foreign language teachers.

Equine and human mutual welfare: a whole subject? Critical aspects and possible strategies in equine-assisted activities and therapies

General aim of the study is equine welfare, particularly concerning different husbandry methodic and inter-specific relational factors. Specific aim is the evaluation of possible mutual (to humans and to equines) benefits and the analysis of critical factors/strength points, of human-horse relationship within Therapeutic Riding context (TR). The peculiarities of human-horse relationship (compared to the bond with “Pet”) are analyzed, concerning their socio-anthropological, psychological, psycho-dynamic distinctive characteristics. 8 European representative therapeutic riding centers (TRC) were therefore selected (on the basis of their different animals’ husbandry criteria, and of the different rehabilitative methodologies adopted). TRC were investigated through 2 different questionnaires, specifically settled to access objective/subjective animal welfare parameters; the quality of human-horse relationship; technicians’ emotional experienced. 3 Centers were further selected, and behavioral (145 hours of behavioral recording) and physiological parameters (heart rate and heart rate variability) were evaluated, aimed to access equine welfare and horses’ adaptive responses/coping (towards general environment and towards TR job). Moreover a specific “handling-task” was ideated and experimented, aimed to measure the quality of TR technicians-horses relationship. We did therefore evaluate both the individual horses’ responses and the possible differences among Centers. Data collected highlight the lack of univocal standardized methodic, concerning the general animals’ management and the specific methodologies (aimed to improve animal welfare and to empower TR efficacy). Some positive and some critical aspects were detected concerning TR personnel-horse relationship. Another experimental approach did evaluate the efficacy (concerning the mutual benefits’ empowerment) of an “ethologically-fitted” TR intervention, aimed to educate children to and through the relationship with horses. Our data evidenced that the improvement of human horse relationship, through structured educational programs for TR personnel might have important consequences both to human and equine welfare.

Social inclusion of vulnerable groups through participatory and emancipatory approaches. Implementing active citizenship and socially innovative actions in the framework of civil & human rights model of disability

The research hypothesis of the thesis is that “an open participation in the co-creation of the services and environments, makes life easier for vulnerable groups”; assuming that the participatory and emancipatory approaches are processes of possible actions and changes aimed at facilitating people’s lives. The adoption of these approaches is put forward as the common denominator of social innovative practices that supporting inclusive processes allow a shift from a medical model to a civil and human rights approach to disability. The theoretical basis of this assumption finds support in many principles of Inclusive Education and the main focus of the hypothesis of research is on participation and emancipation as approaches aimed at facing emerging and existing problems related to inclusion. The framework of reference for the research is represented by the perspectives adopted by several international documents concerning policies and interventions to promote and support the leadership and participation of vulnerable groups. In the first part an in-depth analysis of the main academic publications on the central themes of the thesis has been carried out. After investigating the framework of reference, the analysis focuses on the main tools of participatory and emancipatory approaches, which are able to connect with the concepts of active citizenship and social innovation. In the second part two case studies concerning participatory and emancipatory approaches in the areas of concern are presented and analyzed as example of the improvement of inclusion, through the involvement and participation of persons with disability. The research has been developed using a holistic and interdisciplinary approach, aimed at providing a knowledge-base that fosters a shift from a situation of passivity and care towards a new scenario based on the person’s commitment in the elaboration of his/her own project of life.

Knowledge Management and Decision Support for Sustainable Land Management

Much research has focused on desertification and land degradation assessments without putting sufficient emphasis on prevention and mitigation, although the concept of sustainable land management (SLM) is increasingly being acknowledged. A variety of SLM measures have already been applied at the local level, but they are rarely adequately recognised, evaluated, shared or used for decision support. WOCAT (World Overview of Technologies and Approaches) has developed an internationally recognised, standardised methodology to document and evaluate SLM technologies and approaches, including spatial distribution, allowing the sharing of SLM knowledge worldwide. The recent methodological integration into a participatory process allows now analysing and using this knowledge for decision support at the local and national level. The use of the WOCAT tools stimulates evaluation (self-evaluation as well as learning from comparing experiences) within SLM initiatives where all too often there is not only insufficient monitoring but also a lack of critical analysis. The comprehensive questionnaires and database system facilitate to document, evaluate and disseminate local experiences of SLM technologies and their implementation approaches. This evaluation process - in a team of experts and together with land users - greatly enhances understanding of the reasons behind successful (or failed) local practices. It has now been integrated into a new methodology for appraising and selecting SLM options. The methodology combines a local collective learning and decision approach with the use of the evaluated global best practices from WOCAT in a concise three step process: i) identifying land degradation and locally applied solutions in a stakeholder learning workshop; ii) assessing local solutions with the standardised WOCAT tool; iii) jointly selecting promising strategies for implementation with the help of a decision support tool. The methodology has been implemented in various countries and study sites around the world mainly within the FAO LADA (Land Degradation Assessment Project) and the EU-funded DESIRE project. Investments in SLM must be carefully assessed and planned on the basis of properly documented experiences and evaluated impacts and benefits: concerted efforts are needed and sufficient resources must be mobilised to tap the wealth of knowledge and learn from SLM successes.

Initial clinical experience with the 4-F self-expanding XPERT stent system for infrapopliteal treatment of patients with severe claudication and critical limb ischemia

PURPOSE: To evaluate the primary success and short-term patency associated with a new 4-F sheath-compatible self-expanding nitinol stent after failed conventional angioplasty of distal popliteal and infrapopliteal lesions in severe lifestyle-limiting claudication (LLC) and chronic critical limb ischemia (CLI). MATERIALS AND METHODS: Between May 2003 and July 2005, 35 patients with Rutherford category 3-5 disease (16 patients with CLI, 19 patients with LLC) underwent percutaneous transluminal angioplasty (PTA) and stent implantation. Indications for stent placement were residual stenosis, flow-limiting dissections, or elastic recoil after PTA. Before and after the intervention and during the 6-month follow-up, clinical investigation, color-flow and duplex Doppler ultrasonography, and digital subtraction angiography were performed. Technical success, primary patency at 6 months, clinical improvement as defined by Rutherford with clinical and hemodynamic measures, and complications were evaluated. RESULTS: A total of 22 patients underwent distal popliteal artery stent placement and 13 underwent tibioperoneal artery stent placement. Stent implantation was successfully performed in all patients. After stent placement, the primary cumulative patency rate for the study group at 6 months was 82%. The mean resting ankle-brachial index at baseline was 0.50 +/- 0.16 and significantly increased to 0.90 +/- 0.17 at 12-24 hours after intervention and 0.82 +/- 0.24 at latest follow-up (P < .001 for both). The sustained clinical improvement rate was 80% at the 6-month follow-up. The 6-month limb salvage rate regarding major amputation was 100%. The rate of major complications was 17%. CONCLUSIONS: Infrapopliteal application of the new nitinol stent is a safe, feasible, and effective method with good short-term patency rate in the treatment of severe LLC and chronic CLI.

Photodegradation and photostabilization of weathered wood flour filled polyethylene composites

Wood plastic composites (WPCs) have gained popularity as building materials because of their usefulness in replacing solid wood in a variety of applications. These composites are promoted as being low-maintenance, high-durability products. However, it has been shown that WPCs exposed to weathering may experience a color change and/or loss in mechanical properties. An important requirement for building materials used in outdoor applications is the retention of their aesthetic qualities and mechanical properties during service life. Therefore, it is critical to understand the photodegradation mechanisms of WPCs exposed to UV radiation and to develop approaches to stabilize these composites (both unstabilized and stabilized) as well as the effect of weathering on the color fade and the retention of mechanical properties were characterized. Since different methods of manufacturing WPCs lead to different surface characteristics, which can influence weathering, the effect of manufacturing method on the photodegradation of WPCs was investigated first. Wood flour (WF) filled high-density polyethylene (HDPE) composite samples were either injection molded, extruded, or extruded and then planed. Fourier transform infrared (FTIR) spectroscopy was used to monitor the surface chemistry of the manufactured composites. The spectra showed that the surface of planed samples had more wood component than extruded and injection molded samples, respectively. After weathering, the samples were analyzed for color fade, and loss of flexural properties. The final lightness of the composites was not dependent upon the manufacturing method. However the mechanical property loss was dependent upon manufacturing method. The samples with more wood component at the surface (planed samples) experienced a larger percentage of total loss in flexural properties after weathering due to a greater effect of moisture on the samples. The change in surface chemistry of HDPE and WF/HDPE composites after weathering was studied using spectroscopic techniques. X-ray photoelectron spectroscopy (XPS) was used to characterize the occurrence of surface oxidation whereas FTIR spectroscopy was used to monitor the development of degradation products, such as carbonyl groups and vinyl groups, and to determine changes in HDPE crystallinity. Surface oxidation occurred immediately after exposure for both the neat HDPE and WF/HDPE composites. After weathering, the surface of the WF/HDPE composites was oxidized to a greater extent than the neat HDPE after weathering. This suggests that photodegradation is exacerbated by the addition of the carbonyl functional groups of the wood fibers within the HDPE atrix during composite manufacturing. While neat HDPE may undergo cross-linking in the initial stages of accelerated weathering, the WF may physically hinder the ability of the HDPE to cross-link resulting in the potential for HDPE chain scission to dominate in the initial weathering stages of the WF/HDPE composites. To determine which photostabilizers are most effective for WF/HDPE composites, factorial experimental designes were used to determine the effects of adding two hindered amine light stabilizers, an ultraviolet absorber, and a pigment on the color made and mechanical properties of both unweathered and UV weathered samples. Both the pigment and ultraviolet absorber were more effective photostabilizers for WF/HDPE composites than hinder amine light stabilizers. The ineffectiveness of hindered amine light stabilizers in protecting WPCs against UV radiation was attribuated to the acid/base reactions occurring between the WF and hindered amine light stabilizer. The efficiency of an ultraviolet absorber and/or pigment was also examined by incorporating different concentration of an ultraviolet absorber and/or pigment into WF/HDPE composites. Color change and flexural properties were determined after accelerated UV weathering. The lightness of the composite after weathering was influenced by the concentration of both the ultraviolet absorber by masking the bleaching wood component as well as blocking UV light. Flexural MOE loss was influenced by an increase in ultraviolet absorber concentration, but increasing pigment concentration from 1 to 2% had little influence on MOE loss. However, increasing both ultraviolet absorber and pigment concentration resulted in improved strength properties over the unstabilized composites after 3000 h of weather. Finally, the change in surface chemistry due to weathering of WF/HDPE composites that were either unstabilized or stabilized with an ultraviolet absorber and/or pigment was analyzed using FTIR spectroscopy. The samples were tested for loss in modulus of elasticity, carbonyl and vinyl group formation at the surface, and change in HDPE crystallinity. It was concluded that structural changes in the samples; carbonyl group formation, terminal vinyl group formation, and crystallinity changes cannot reliably be used to predict changes in modulus of elasticity using a simple linear relationship. The effect of cross-linking, chain scission, and crystallinity changes due to ultraviolet exposure as well as the interfacial degradation due to moisture exposure are inter-related factors when weathering HDPE and WF/HDPE composites.

Modeling and classifying variable width riparian zones utilizing digital elevation models, flood height data, digital soil data and national wetlands inventory : a new approach for riparian zone delineation

Riparian zones are dynamic, transitional ecosystems between aquatic and terrestrial ecosystems with well defined vegetation and soil characteristics. Development of an all-encompassing definition for riparian ecotones, because of their high variability, is challenging. However, there are two primary factors that all riparian ecotones are dependent on: the watercourse and its associated floodplain. Previous approaches to riparian boundary delineation have utilized fixed width buffers, but this methodology has proven to be inadequate as it only takes the watercourse into consideration and ignores critical geomorphology, associated vegetation and soil characteristics. Our approach offers advantages over other previously used methods by utilizing: the geospatial modeling capabilities of ArcMap GIS; a better sampling technique along the water course that can distinguish the 50-year flood plain, which is the optimal hydrologic descriptor of riparian ecotones; the Soil Survey Database (SSURGO) and National Wetland Inventory (NWI) databases to distinguish contiguous areas beyond the 50-year plain; and land use/cover characteristics associated with the delineated riparian zones. The model utilizes spatial data readily available from Federal and State agencies and geospatial clearinghouses. An accuracy assessment was performed to assess the impact of varying the 50-year flood height, changing the DEM spatial resolution (1, 3, 5 and 10m), and positional inaccuracies with the National Hydrography Dataset (NHD) streams layer on the boundary placement of the delineated variable width riparian ecotones area. The result of this study is a robust and automated GIS based model attached to ESRI ArcMap software to delineate and classify variable-width riparian ecotones.

Critical evaluation of diameter increment modelling in the Great Lakes region

Simulations of forest stand dynamics in a modelling framework including Forest Vegetation Simulator (FVS) are diameter driven, thus the diameter or basal area increment model needs a special attention. This dissertation critically evaluates diameter or basal area increment models and modelling approaches in the context of the Great Lakes region of the United States and Canada. A set of related studies are presented that critically evaluate the sub-model for change in individual tree basal diameter used in the Forest Vegetation Simulator (FVS), a dominant forestry model in the Great Lakes region. Various historical implementations of the STEMS (Stand and Tree Evaluation and Modeling System) family of diameter increment models, including the current public release of the Lake States variant of FVS (LS-FVS), were tested for the 30 most common tree species using data from the Michigan Forest Inventory and Analysis (FIA) program. The results showed that current public release of the LS-FVS diameter increment model over-predicts 10-year diameter increment by 17% on average. Also the study affirms that a simple adjustment factor as a function of a single predictor, dbh (diameter at breast height) used in the past versions, provides an inadequate correction of model prediction bias. In order to re-engineer the basal diameter increment model, the historical, conceptual and philosophical differences among the individual tree increment model families and their modelling approaches were analyzed and discussed. Two underlying conceptual approaches toward diameter or basal area increment modelling have been often used: the potential-modifier (POTMOD) and composite (COMP) approaches, which are exemplified by the STEMS/TWIGS and Prognosis models, respectively. It is argued that both approaches essentially use a similar base function and neither is conceptually different from a biological perspective, even though they look different in their model forms. No matter what modelling approach is used, the base function is the foundation of an increment model. Two base functions – gamma and Box-Lucas – were identified as candidate base functions for forestry applications. The results of a comparative analysis of empirical fits showed that quality of fit is essentially similar, and both are sufficiently detailed and flexible for forestry applications. The choice of either base function in order to model diameter or basal area increment is dependent upon personal preference; however, the gamma base function may be preferred over the Box-Lucas, as it fits the periodic increment data in both a linear and nonlinear composite model form. Finally, the utility of site index as a predictor variable has been criticized, as it has been widely used in models for complex, mixed species forest stands though not well suited for this purpose. An alternative to site index in an increment model was explored, using site index and a combination of climate variables and Forest Ecosystem Classification (FEC) ecosites and data from the Province of Ontario, Canada. The results showed that a combination of climate and FEC ecosites variables can replace site index in the diameter increment model.

Assessing the state of Sustainable Land Management research in Kyrgyzstan and Tajikistan

This article synthesizes findings from a review of the state of research on sustainable land management in Kyrgyzstan and Tajikistan and from an analysis of the interface between research and action. Using the Global Land Project (GLP 2005) analytical framework, we analyzed the distribution of 131 selected publications (including a clearly defined set of local and international academic and gray literature) across the framework's components and links in a social–ecological system. There is a strong emphasis in the literature on the impact of changes in land use and management on ecosystems; however, there is little research on the implications for ecosystem services. This finding is opposed to that of a similar analysis of publications at the global scale (Björnsen Gurung et al 2012). Another major gap was the lack of research on Kyrgyzstan and Tajikistan regarding the influence of global factors on social and ecological systems, despite social, economic, and political integration into global structures since the collapse of the Soviet Union and the increasing influence of climate change. Our analysis disaggregated academic literature published in the region and international academic literature, revealing stark differences. These differences are partly attributable to the legacy of the late Soviet era principle of “rational use of land resources,” which fit the planned economy but lacks approaches for decentralized resource governance. Finally, the emphasis of research on systems knowledge, the lack of transdisciplinary research, and the critical feedback of stakeholders at a regional sustainable land management forum suggest that actionable sustainable land management research on Kyrgyzstan and Tajikistan is rare. Recommendations are made for targeted, application-focused, multistakeholder research and knowledge sharing, including local and international researchers as well as practitioners, policy-makers, and land users.

A novel phosphorylation-independent interaction between SMG6 and UPF1 is essential for human NMD

Eukaryotic mRNAs with premature translation-termination codons (PTCs) are recognized and eliminated by nonsense-mediated mRNA decay (NMD). NMD substrates can be degraded by different routes that all require phosphorylated UPF1 (P-UPF1) as a starting point. The endonuclease SMG6, which cleaves mRNA near the PTC, is one of the three known NMD factors thought to be recruited to nonsense mRNAs via an interaction with P-UPF1, leading to eventual mRNA degradation. By artificial tethering of SMG6 and mutants thereof to a reporter mRNA combined with knockdowns of various NMD factors, we demonstrate that besides its endonucleolytic activity, SMG6 also requires UPF1 and SMG1 to reduce reporter mRNA levels. Using in vivo and in vitro approaches, we further document that SMG6 and the unique stalk region of the UPF1 helicase domain, along with a contribution from the SQ domain, form a novel interaction and we also show that this region of the UPF1 helicase domain is critical for SMG6 function and NMD. Our results show that this interaction is required for NMD and for the capability of tethered SMG6 to degrade its bound RNA, suggesting that it contributes to the intricate regulation of UPF1 and SMG6 enzymatic activities.

The Quiet Eye in Motor Performance and Learning – Functional and Theoretical Considerations

High precision in motor skill performance, in both sport and other domains (e.g. surgery and aviation), requires the efficient coupling of perceptual inputs (e.g. vision) and motor actions. A particular gaze strategy, which has received much attention within the literature, has been shown to predict both inter- (expert vs. novice) and intra-individual (successful vs. unsuccessful) motor performance (see Vine et al., 2014). Vickers (1996) labelled this phenomenon the quiet eye (QE) which is defined as the final fixation before the initiation of the crucial phase of movement. While the positive influence of a long QE on accuracy has been revealed in a range of different motor skills, there is a growing number of studies suggesting that the relationship between QE and motor performance is not entirely monotonic. This raises interesting questions regarding the QE’s purview, and the theoretical approaches explaining its functionality. This talk aims to present an overview of the issues described above, and to discuss contemporary research and experimental approaches to examining the QE phenomenon. In the first part of the talk Dr. Vine will provide a brief and critical review of the literature, highlighting recent empirical advancements and potential directions for future research. In the second part, Dr. Klostermann will communicate three different theoretical approaches to explain the relationship between QE and motor performance. Drawing upon aspects of all three of these theoretical approaches, a functional inhibition role for the QE (related to movement parameterisation) will be proposed.

Assessment of CTF boiling transition and critical heat flux modeling capabilities using the OECD/NRC BFBT and PSBT benchmark databases

The need to refine models for best-estimate calculations, based on good-quality experimental data, has been expressed in many recent meetings in the field of nuclear applications. The modeling needs arising in this respect should not be limited to the currently available macroscopic methods but should be extended to next-generation analysis techniques that focus on more microscopic processes. One of the most valuable databases identified for the thermalhydraulics modeling was developed by the Nuclear Power Engineering Corporation (NUPEC), Japan. From 1987 to 1995, NUPEC performed steady-state and transient critical power and departure from nucleate boiling (DNB) test series based on the equivalent full-size mock-ups. Considering the reliability not only of the measured data, but also other relevant parameters such as the system pressure, inlet sub-cooling and rod surface temperature, these test series supplied the first substantial database for the development of truly mechanistic and consistent models for boiling transition and critical heat flux. Over the last few years the Pennsylvania State University (PSU) under the sponsorship of the U.S. Nuclear Regulatory Commission (NRC) has prepared, organized, conducted and summarized the OECD/NRC Full-size Fine-mesh Bundle Tests (BFBT) Benchmark. The international benchmark activities have been conducted in cooperation with the Nuclear Energy Agency/Organization for Economic Co-operation and Development (NEA/OECD) and Japan Nuclear Energy Safety (JNES) organization, Japan. Consequently, the JNES has made available the Boiling Water Reactor (BWR) NUPEC database for the purposes of the benchmark. Based on the success of the OECD/NRC BFBT benchmark the JNES has decided to release also the data based on the NUPEC Pressurized Water Reactor (PWR) subchannel and bundle tests for another follow-up international benchmark entitled OECD/NRC PWR Subchannel and Bundle Tests (PSBT) benchmark. This paper presents an application of the joint Penn State University/Technical University of Madrid (UPM) version of the well-known subchannel code COBRA-TF, namely CTF, to the critical power and departure from nucleate boiling (DNB) exercises of the OECD/NRC BFBT and PSBT benchmarks

Contributions to Speech Analytics based on Speech Recognition and Topic Identification

La última década ha sido testigo de importantes avances en el campo de la tecnología de reconocimiento de voz. Los sistemas comerciales existentes actualmente poseen la capacidad de reconocer habla continua de múltiples locutores, consiguiendo valores aceptables de error, y sin la necesidad de realizar procedimientos explícitos de adaptación. A pesar del buen momento que vive esta tecnología, el reconocimiento de voz dista de ser un problema resuelto. La mayoría de estos sistemas de reconocimiento se ajustan a dominios particulares y su eficacia depende de manera significativa, entre otros muchos aspectos, de la similitud que exista entre el modelo de lenguaje utilizado y la tarea específica para la cual se está empleando. Esta dependencia cobra aún más importancia en aquellos escenarios en los cuales las propiedades estadísticas del lenguaje varían a lo largo del tiempo, como por ejemplo, en dominios de aplicación que involucren habla espontánea y múltiples temáticas. En los últimos años se ha evidenciado un constante esfuerzo por mejorar los sistemas de reconocimiento para tales dominios. Esto se ha hecho, entre otros muchos enfoques, a través de técnicas automáticas de adaptación. Estas técnicas son aplicadas a sistemas ya existentes, dado que exportar el sistema a una nueva tarea o dominio puede requerir tiempo a la vez que resultar costoso. Las técnicas de adaptación requieren fuentes adicionales de información, y en este sentido, el lenguaje hablado puede aportar algunas de ellas. El habla no sólo transmite un mensaje, también transmite información acerca del contexto en el cual se desarrolla la comunicación hablada (e.g. acerca del tema sobre el cual se está hablando). Por tanto, cuando nos comunicamos a través del habla, es posible identificar los elementos del lenguaje que caracterizan el contexto, y al mismo tiempo, rastrear los cambios que ocurren en estos elementos a lo largo del tiempo. Esta información podría ser capturada y aprovechada por medio de técnicas de recuperación de información (information retrieval) y de aprendizaje de máquina (machine learning). Esto podría permitirnos, dentro del desarrollo de mejores sistemas automáticos de reconocimiento de voz, mejorar la adaptación de modelos del lenguaje a las condiciones del contexto, y por tanto, robustecer al sistema de reconocimiento en dominios con condiciones variables (tales como variaciones potenciales en el vocabulario, el estilo y la temática). En este sentido, la principal contribución de esta Tesis es la propuesta y evaluación de un marco de contextualización motivado por el análisis temático y basado en la adaptación dinámica y no supervisada de modelos de lenguaje para el robustecimiento de un sistema automático de reconocimiento de voz. Esta adaptación toma como base distintos enfoque de los sistemas mencionados (de recuperación de información y aprendizaje de máquina) mediante los cuales buscamos identificar las temáticas sobre las cuales se está hablando en una grabación de audio. Dicha identificación, por lo tanto, permite realizar una adaptación del modelo de lenguaje de acuerdo a las condiciones del contexto. El marco de contextualización propuesto se puede dividir en dos sistemas principales: un sistema de identificación de temática y un sistema de adaptación dinámica de modelos de lenguaje. Esta Tesis puede describirse en detalle desde la perspectiva de las contribuciones particulares realizadas en cada uno de los campos que componen el marco propuesto: _ En lo referente al sistema de identificación de temática, nos hemos enfocado en aportar mejoras a las técnicas de pre-procesamiento de documentos, asimismo en contribuir a la definición de criterios más robustos para la selección de index-terms. – La eficiencia de los sistemas basados tanto en técnicas de recuperación de información como en técnicas de aprendizaje de máquina, y específicamente de aquellos sistemas que particularizan en la tarea de identificación de temática, depende, en gran medida, de los mecanismos de preprocesamiento que se aplican a los documentos. Entre las múltiples operaciones que hacen parte de un esquema de preprocesamiento, la selección adecuada de los términos de indexado (index-terms) es crucial para establecer relaciones semánticas y conceptuales entre los términos y los documentos. Este proceso también puede verse afectado, o bien por una mala elección de stopwords, o bien por la falta de precisión en la definición de reglas de lematización. En este sentido, en este trabajo comparamos y evaluamos diferentes criterios para el preprocesamiento de los documentos, así como también distintas estrategias para la selección de los index-terms. Esto nos permite no sólo reducir el tamaño de la estructura de indexación, sino también mejorar el proceso de identificación de temática. – Uno de los aspectos más importantes en cuanto al rendimiento de los sistemas de identificación de temática es la asignación de diferentes pesos a los términos de acuerdo a su contribución al contenido del documento. En este trabajo evaluamos y proponemos enfoques alternativos a los esquemas tradicionales de ponderado de términos (tales como tf-idf ) que nos permitan mejorar la especificidad de los términos, así como también discriminar mejor las temáticas de los documentos. _ Respecto a la adaptación dinámica de modelos de lenguaje, hemos dividimos el proceso de contextualización en varios pasos. – Para la generación de modelos de lenguaje basados en temática, proponemos dos tipos de enfoques: un enfoque supervisado y un enfoque no supervisado. En el primero de ellos nos basamos en las etiquetas de temática que originalmente acompañan a los documentos del corpus que empleamos. A partir de estas, agrupamos los documentos que forman parte de la misma temática y generamos modelos de lenguaje a partir de dichos grupos. Sin embargo, uno de los objetivos que se persigue en esta Tesis es evaluar si el uso de estas etiquetas para la generación de modelos es óptimo en términos del rendimiento del reconocedor. Por esta razón, nosotros proponemos un segundo enfoque, un enfoque no supervisado, en el cual el objetivo es agrupar, automáticamente, los documentos en clusters temáticos, basándonos en la similaridad semántica existente entre los documentos. Por medio de enfoques de agrupamiento conseguimos mejorar la cohesión conceptual y semántica en cada uno de los clusters, lo que a su vez nos permitió refinar los modelos de lenguaje basados en temática y mejorar el rendimiento del sistema de reconocimiento. – Desarrollamos diversas estrategias para generar un modelo de lenguaje dependiente del contexto. Nuestro objetivo es que este modelo refleje el contexto semántico del habla, i.e. las temáticas más relevantes que se están discutiendo. Este modelo es generado por medio de la interpolación lineal entre aquellos modelos de lenguaje basados en temática que estén relacionados con las temáticas más relevantes. La estimación de los pesos de interpolación está basada principalmente en el resultado del proceso de identificación de temática. – Finalmente, proponemos una metodología para la adaptación dinámica de un modelo de lenguaje general. El proceso de adaptación tiene en cuenta no sólo al modelo dependiente del contexto sino también a la información entregada por el proceso de identificación de temática. El esquema usado para la adaptación es una interpolación lineal entre el modelo general y el modelo dependiente de contexto. Estudiamos también diferentes enfoques para determinar los pesos de interpolación entre ambos modelos. Una vez definida la base teórica de nuestro marco de contextualización, proponemos su aplicación dentro de un sistema automático de reconocimiento de voz. Para esto, nos enfocamos en dos aspectos: la contextualización de los modelos de lenguaje empleados por el sistema y la incorporación de información semántica en el proceso de adaptación basado en temática. En esta Tesis proponemos un marco experimental basado en una arquitectura de reconocimiento en ‘dos etapas’. En la primera etapa, empleamos sistemas basados en técnicas de recuperación de información y aprendizaje de máquina para identificar las temáticas sobre las cuales se habla en una transcripción de un segmento de audio. Esta transcripción es generada por el sistema de reconocimiento empleando un modelo de lenguaje general. De acuerdo con la relevancia de las temáticas que han sido identificadas, se lleva a cabo la adaptación dinámica del modelo de lenguaje. En la segunda etapa de la arquitectura de reconocimiento, usamos este modelo adaptado para realizar de nuevo el reconocimiento del segmento de audio. Para determinar los beneficios del marco de trabajo propuesto, llevamos a cabo la evaluación de cada uno de los sistemas principales previamente mencionados. Esta evaluación es realizada sobre discursos en el dominio de la política usando la base de datos EPPS (European Parliamentary Plenary Sessions - Sesiones Plenarias del Parlamento Europeo) del proyecto europeo TC-STAR. Analizamos distintas métricas acerca del rendimiento de los sistemas y evaluamos las mejoras propuestas con respecto a los sistemas de referencia. ABSTRACT The last decade has witnessed major advances in speech recognition technology. Today’s commercial systems are able to recognize continuous speech from numerous speakers, with acceptable levels of error and without the need for an explicit adaptation procedure. Despite this progress, speech recognition is far from being a solved problem. Most of these systems are adjusted to a particular domain and their efficacy depends significantly, among many other aspects, on the similarity between the language model used and the task that is being addressed. This dependence is even more important in scenarios where the statistical properties of the language fluctuates throughout the time, for example, in application domains involving spontaneous and multitopic speech. Over the last years there has been an increasing effort in enhancing the speech recognition systems for such domains. This has been done, among other approaches, by means of techniques of automatic adaptation. These techniques are applied to the existing systems, specially since exporting the system to a new task or domain may be both time-consuming and expensive. Adaptation techniques require additional sources of information, and the spoken language could provide some of them. It must be considered that speech not only conveys a message, it also provides information on the context in which the spoken communication takes place (e.g. on the subject on which it is being talked about). Therefore, when we communicate through speech, it could be feasible to identify the elements of the language that characterize the context, and at the same time, to track the changes that occur in those elements over time. This information can be extracted and exploited through techniques of information retrieval and machine learning. This allows us, within the development of more robust speech recognition systems, to enhance the adaptation of language models to the conditions of the context, thus strengthening the recognition system for domains under changing conditions (such as potential variations in vocabulary, style and topic). In this sense, the main contribution of this Thesis is the proposal and evaluation of a framework of topic-motivated contextualization based on the dynamic and non-supervised adaptation of language models for the enhancement of an automatic speech recognition system. This adaptation is based on an combined approach (from the perspective of both information retrieval and machine learning fields) whereby we identify the topics that are being discussed in an audio recording. The topic identification, therefore, enables the system to perform an adaptation of the language model according to the contextual conditions. The proposed framework can be divided in two major systems: a topic identification system and a dynamic language model adaptation system. This Thesis can be outlined from the perspective of the particular contributions made in each of the fields that composes the proposed framework: _ Regarding the topic identification system, we have focused on the enhancement of the document preprocessing techniques in addition to contributing in the definition of more robust criteria for the selection of index-terms. – Within both information retrieval and machine learning based approaches, the efficiency of topic identification systems, depends, to a large extent, on the mechanisms of preprocessing applied to the documents. Among the many operations that encloses the preprocessing procedures, an adequate selection of index-terms is critical to establish conceptual and semantic relationships between terms and documents. This process might also be weakened by a poor choice of stopwords or lack of precision in defining stemming rules. In this regard we compare and evaluate different criteria for preprocessing the documents, as well as for improving the selection of the index-terms. This allows us to not only reduce the size of the indexing structure but also to strengthen the topic identification process. – One of the most crucial aspects, in relation to the performance of topic identification systems, is to assign different weights to different terms depending on their contribution to the content of the document. In this sense we evaluate and propose alternative approaches to traditional weighting schemes (such as tf-idf ) that allow us to improve the specificity of terms, and to better identify the topics that are related to documents. _ Regarding the dynamic language model adaptation, we divide the contextualization process into different steps. – We propose supervised and unsupervised approaches for the generation of topic-based language models. The first of them is intended to generate topic-based language models by grouping the documents, in the training set, according to the original topic labels of the corpus. Nevertheless, a goal of this Thesis is to evaluate whether or not the use of these labels to generate language models is optimal in terms of recognition accuracy. For this reason, we propose a second approach, an unsupervised one, in which the objective is to group the data in the training set into automatic topic clusters based on the semantic similarity between the documents. By means of clustering approaches we expect to obtain a more cohesive association of the documents that are related by similar concepts, thus improving the coverage of the topic-based language models and enhancing the performance of the recognition system. – We develop various strategies in order to create a context-dependent language model. Our aim is that this model reflects the semantic context of the current utterance, i.e. the most relevant topics that are being discussed. This model is generated by means of a linear interpolation between the topic-based language models related to the most relevant topics. The estimation of the interpolation weights is based mainly on the outcome of the topic identification process. – Finally, we propose a methodology for the dynamic adaptation of a background language model. The adaptation process takes into account the context-dependent model as well as the information provided by the topic identification process. The scheme used for the adaptation is a linear interpolation between the background model and the context-dependent one. We also study different approaches to determine the interpolation weights used in this adaptation scheme. Once we defined the basis of our topic-motivated contextualization framework, we propose its application into an automatic speech recognition system. We focus on two aspects: the contextualization of the language models used by the system, and the incorporation of semantic-related information into a topic-based adaptation process. To achieve this, we propose an experimental framework based in ‘a two stages’ recognition architecture. In the first stage of the architecture, Information Retrieval and Machine Learning techniques are used to identify the topics in a transcription of an audio segment. This transcription is generated by the recognition system using a background language model. According to the confidence on the topics that have been identified, the dynamic language model adaptation is carried out. In the second stage of the recognition architecture, an adapted language model is used to re-decode the utterance. To test the benefits of the proposed framework, we carry out the evaluation of each of the major systems aforementioned. The evaluation is conducted on speeches of political domain using the EPPS (European Parliamentary Plenary Sessions) database from the European TC-STAR project. We analyse several performance metrics that allow us to compare the improvements of the proposed systems against the baseline ones.