Mutual Information and Perplexity based Clustering of Dialogue Information for Dynamic Adaptation of Language Models

We present two approaches to cluster dialogue-based information obtained by the speech understanding module and the dialogue manager of a spoken dialogue system. The purpose is to estimate a language model related to each cluster, and use them to dynamically modify the model of the speech recognizer at each dialogue turn. In the first approach we build the cluster tree using local decisions based on a Maximum Normalized Mutual Information criterion. In the second one we take global decisions, based on the optimization of the global perplexity of the combination of the cluster-related LMs. Our experiments show a relative reduction of the word error rate of 15.17%, which helps to improve the performance of the understanding and the dialogue manager modules.

Mutual Information and Perplexity based Clustering of Dialogue Information for Dynamic Adaptation of Language Models

We present two approaches to cluster dialogue-based information obtained by the speech understanding module and the dialogue manager of a spoken dialogue system. The purpose is to estimate a language model related to each cluster, and use them to dynamically modify the model of the speech recognizer at each dialogue turn. In the first approach we build the cluster tree using local decisions based on a Maximum Normalized Mutual Information criterion. In the second one we take global decisions, based on the optimization of the global perplexity of the combination of the cluster-related LMs. Our experiments show a relative reduction of the word error rate of 15.17%, which helps to improve the performance of the understanding and the dialogue manager modules.

Validation tool for smart screening of language disorders in pediatrics care

Primary-care pediatricians could play a key role in early detection of development disorders as quick as they might have enough time and knowledge for suitable screenings at clinical routine. This research paper focuses on the development and validation of a knowledge-based web tool whose aim is to support a smart detection of developmental disorders in early childhood. Thus, the use of the system can trigger the necessary preventive and therapeutic actions from birth until the age of six. The platform was designed on the basis of an analysis of significant 21 cases of children with language disorders that supported the creation of a specific knowledge base, its ontology and a set of description logic relations. The resulting system is being validated in a scalable approach with a team of seven experts from the fields of neonathology, pediatrics, neurology and language therapy.

Development of a Knowledge Base for smart screening of language disorders in primary care

Monitoring of neuro-evolutive development from birth until the age of six is a decisive factor in a child's quality of life. Early detection of development disorders in early childhood can facilitate necessary diagnosis and/or treatment. Primary-care pediatricians play a key role in early detection of development alterations as they can undertake the preventive and therapeutic actions necessary in the interest of a child's optimal development. The focus of this research paper is the construction of a Knowledge Base for smart screening aimed to assist pediatricians in processes of early referral in language disorders. The proposed model provides health professionals with a decision-making tool that supports referral processes. In this way, essential diagnostic and/or therapeutic actions are triggered for a comprehensive individual development. The resulting system was developed on the basis of an analysis and verification of 21 cases of children with language disorders.

On the dynamic adaptation of language models based on dialogue information

We present an approach to adapt dynamically the language models (LMs) used by a speech recognizer that is part of a spoken dialogue system. We have developed a grammar generation strategy that automatically adapts the LMs using the semantic information that the user provides (represented as dialogue concepts), together with the information regarding the intentions of the speaker (inferred by the dialogue manager, and represented as dialogue goals). We carry out the adaptation as a linear interpolation between a background LM, and one or more of the LMs associated to the dialogue elements (concepts or goals) addressed by the user. The interpolation weights between those models are automatically estimated on each dialogue turn, using measures such as the posterior probabilities of concepts and goals, estimated as part of the inference procedure to determine the actions to be carried out. We propose two approaches to handle the LMs related to concepts and goals. Whereas in the first one we estimate a LM for each one of them, in the second one we apply several clustering strategies to group together those elements that share some common properties, and estimate a LM for each cluster. Our evaluation shows how the system can estimate a dynamic model adapted to each dialogue turn, which helps to improve the performance of the speech recognition (up to a 14.82% of relative improvement), which leads to an improvement in both the language understanding and the dialogue management tasks.

Deployment and validation of a smart system for screening of language disorders in primary care

Neuro-evolutive development from birth until the age of six years is a decisive factor in a child?s quality of life. Early detection of development disorders in early childhood can facilitate necessary diagnosis and/or treatment. Primary-care pediatricians play a key role in its detection as they can undertake the preventive and therapeutic actions requested to promote a child?s optimal development. However, the lack of time and little specific knowledge at primary-care avoid to applying continuous early-detection anomalies procedures. This research paper focuses on the deployment and evaluation of a smart system that enhances the screening of language disorders in primary care. Pediatricians get support to proceed with early referral of language disorders. The proposed model provides them with a decision-support tool for referral actions to trigger essential diagnostic and/or therapeutic actions for a comprehensive individual development. The research was conducted by starting from a sample of 60 cases of children with language disorders. Validation was carried out through two complementary steps: first, by including a team of seven experts from the fields of neonatology, pediatrics, neurology and language therapy, and, second, through the evaluation of 21 more previously diagnosed cases. The results obtained show that therapist positively accepted the system proposal in 18 cases (86%) and suggested system redesign for single referral to a speech therapist in three remaining cases.

Large-scale assessment of language proficiency: theoretical and pedagogical reflections on the use of multiple-choice tests

The new requirement placed on students in tertiary settings in Spain to demonstrate a B1 or a B2 proficiency level of English, in accordance with the Common European Framework of Reference for Languages (CEFRL), has led most Spanish universities to develop a program of certification or accreditation of the required level. The first part of this paper aims to provide a rationale for the type of test that has been developed at the Universidad Politécnica de Madrid for the accreditation of a B2 level, a multiple choice version, and to describe how it was constructed and validated. Then, in the second part of the paper, the results from its application to 924 students enrolled in different degree courses at a variety of schools and faculties at the university are analyzed based on a final test version item analysis. To conclude, some theoretical as well as practical conclusions about testing grammar that affect the teaching and learning process are drawn. RESUMEN. Las nuevas exigencias sobre niveles de competencia B1 y B2 en inglés según el Marco Común Europeo de Referencia para las Lenguas (MCERL) que se imponen sobre los estudiantes de grado y posgrado han llevado a la mayoría de las universidades españolas a desarrollar programas de acreditación o de certificación de estos niveles. La primera parte de este trabajo trata sobre las razones que fundamentan la elección de un tipo concreto de examen para la acreditación del nivel B2 de lengua inglesa en la Universidad Politécnica de Madrid. Se trata de un test de opción múltiple y en esta parte del trabajo se describe cómo fue diseñado y validado. En la segunda parte, se analizan los resultados de la aplicación del test a gran escala a un total de 924 estudiantes matriculados en varias escuelas y Facultades de la Universidad. Para terminar, se apuntan una serie de conclusiones teóricas y prácticas sobre la evaluación de la gramática y de qué modo influye en los procesos de enseñanza y aprendizaje.

A supervised cooperative learning system for early detection of language disorders

The Quality of Life of a person may depend on early attention to his neurodevel-opment disorders in childhood. Identification of language disorders under the age of six years old can speed up required diagnosis and/or treatment processes. This paper details the enhancement of a Clinical Decision Support System (CDSS) aimed to assist pediatricians and language therapists at early identification and re-ferral of language disorders. The system helps to fine tune the Knowledge Base of Language Delays (KBLD) that was already developed and validated in clinical routine with 146 children. Medical experts supported the construction of Gades CDSS by getting scientific consensus from literature and fifteen years of regis-tered use cases of children with language disorders. The current research focuses on an innovative cooperative model that allows the evolution of the KBLD of Gades through the supervised evaluation of the CDSS learnings with experts¿ feedback. The deployment of the resulting system is being assessed under a mul-tidisciplinary team of seven experts from the fields of speech therapist, neonatol-ogy, pediatrics, and neurology.

Dynamic topic-based adaptation of language models: a comparison between different approaches

This paper presents a dynamic LM adaptation based on the topic that has been identified on a speech segment. We use LSA and the given topic labels in the training dataset to obtain and use the topic models. We propose a dynamic language model adaptation to improve the recognition performance in "a two stages" AST system. The final stage makes use of the topic identification with two variants: the first on uses just the most probable topic and the other one depends on the relative distances of the topics that have been identified. We perform the adaptation of the LM as a linear interpolation between a background model and topic-based LM. The interpolation weight id dynamically adapted according to different parameters. The proposed method is evaluated on the Spanish partition of the EPPS speech database. We achieved a relative reduction in WER of 11.13% over the baseline system which uses a single blackground LM.

Linked Data for sharing, discovery and re-use of Language Resources at a Web scale

Conferencia por invitación, impartida el 31 d mayo de 2014 en el Workshop on Language Technology Service Platforms: Synergies, Standards, Sharing at LREC2014

Intermittent pluri-sink model and the emergence of complex heterogeneity patterns: a simple paradigm for explaining complexity in soil chemical distributions

The spatial complexity of the distribution of organic matter, chemicals, nutrients, pollutants has been demonstrated to have multifractal nature (Kravchenco et al. [1]). This fact supports the possibility of existence of some emergent heterogeneity structure built under the evolution of the system. The aim of this note is providing a consistent explanation to the mentioned results via an extremely simple model.

Auto-Emergencia de Comunicación Sintáctica en Entornos Estáticos y Dinámicos para Grupos de Robots mediante Evolución Gramatical y Aprendizaje por Refuerzo

La idea de dotar a un grupo de robots o agentes artificiales de un lenguaje ha sido objeto de intenso estudio en las ultimas décadas. Como no podía ser de otra forma los primeros intentos se enfocaron hacia el estudio de la emergencia de vocabularios compartidos convencionalmente por el grupo de robots. Las ventajas que puede ofrecer un léxico común son evidentes, como también lo es que un lenguaje con una estructura más compleja, en la que se pudieran combinar palabras, sería todavía más beneficioso. Surgen así algunas propuestas enfocadas hacia la emergencia de un lenguaje consensuado que muestre una estructura sintáctica similar al lenguaje humano, entre las que se encuentra este trabajo. Tomar el lenguaje humano como modelo supone adoptar algunas de las hipótesis y teorías que disciplinas como la filosofía, la psicología o la lingüística entre otras se han encargado de proponer. Según estas aproximaciones teóricas el lenguaje presenta una doble dimension formal y funcional. En base a su dimensión formal parece claro que el lenguaje sigue unas reglas, por lo que el uso de una gramática se ha considerado esencial para su representación, pero también porque las gramáticas son un dispositivo muy sencillo y potente que permite generar fácilmente estructuras simbólicas. En cuanto a la dimension funcional se ha tenido en cuenta la teoría quizá más influyente de los últimos tiempos, que no es otra que la Teoría de los Actos del Habla. Esta teoría se basa en la idea de Wittgenstein por la que el significado reside en el uso del lenguaje, hasta el punto de que éste se entiende como una manera de actuar y de comportarse, en definitiva como una forma de vida. Teniendo presentes estas premisas en esta tesis se pretende experimentar con modelos computacionales que permitan a un grupo de robots alcanzar un lenguaje común de manera autónoma, simplemente mediante interacciones individuales entre los robots, en forma de juegos de lenguaje. Para ello se proponen tres modelos distintos de lenguaje: • Un modelo basado en gramáticas probabilísticas y aprendizaje por refuerzo en el que las interacciones y el uso del lenguaje son claves para su emergencia y que emplea una gramática generativa estática y diseñada de antemano. Este modelo se aplica a dos grupos distintos: uno formado exclusivamente por robots y otro que combina robots y un humano, de manera que en este segundo caso se plantea un aprendizaje supervisado por humanos. • Un modelo basado en evolución gramatical que permite estudiar no solo el consenso sintáctico, sino también cuestiones relativas a la génesis del lenguaje y que emplea una gramática universal a partir de la cual los robots pueden evolucionar por sí mismos la gramática más apropiada según la situación lingüística que traten en cada momento. • Un modelo basado en evolución gramatical y aprendizaje por refuerzo que toma aspectos de los anteriores y amplia las posibilidades de los robots al permitir desarrollar un lenguaje que se adapta a situaciones lingüísticas dinámicas que pueden cambiar en el tiempo y también posibilita la imposición de restricciones de orden muy frecuentes en las estructuras sintácticas complejas. Todos los modelos implican un planteamiento descentralizado y auto-organizado, de manera que ninguno de los robots es el dueño del lenguaje y todos deben cooperar y colaborar de forma coordinada para lograr el consenso sintáctico. En cada caso se plantean experimentos que tienen como objetivo validar los modelos propuestos, tanto en lo relativo al éxito en la emergencia del lenguaje como en lo relacionado con cuestiones paralelas de importancia, como la interacción hombre-máquina o la propia génesis del lenguaje. ABSTRACT The idea of giving a language to a group of robots or artificial agents has been the subject of intense study in recent decades. The first attempts have focused on the development and emergence of a conventionally shared vocabulary. The advantages that can provide a common vocabulary are evident and therefore a more complex language that combines words would be even more beneficial. Thus some proposals are put forward towards the emergence of a consensual language with a sintactical structure in similar terms to the human language. This work follows this trend. Taking the human language as a model means taking some of the assumptions and theories that disciplines such as philosophy, psychology or linguistics among others have provided. According to these theoretical positions language has a double formal and functional dimension. Based on its formal dimension it seems clear that language follows rules, so that the use of a grammar has been considered essential for representation, but also because grammars are a very simple and powerful device that easily generates these symbolic structures. As for the functional dimension perhaps the most influential theory of recent times, the Theory of Speech Acts has been taken into account. This theory is based on the Wittgenstein’s idea about that the meaning lies in the use of language, to the extent that it is understood as a way of acting and behaving. Having into account these issues this work implements some computational models in order to test if they allow a group of robots to reach in an autonomous way a shared language by means of individual interaction among them, that is by means of language games. Specifically, three different models of language for robots are proposed: • A reinforcement learning based model in which interactions and language use are key to its emergence. This model uses a static probabilistic generative grammar which is designed beforehand. The model is applied to two different groups: one formed exclusively by robots and other combining robots and a human. Therefore, in the second case the learning process is supervised by the human. • A model based on grammatical evolution that allows us to study not only the syntactic consensus, but also the very genesis of language. This model uses a universal grammar that allows robots to evolve for themselves the most appropriate grammar according to the current linguistic situation they deal with. • A model based on grammatical evolution and reinforcement learning that takes aspects of the previous models and increases their possibilities. This model allows robots to develop a language in order to adapt to dynamic language situations that can change over time and also allows the imposition of syntactical order restrictions which are very common in complex syntactic structures. All models involve a decentralized and self-organized approach so that none of the robots is the language’s owner and everyone must cooperate and work together in a coordinated manner to achieve syntactic consensus. In each case experiments are presented in order to validate the proposed models, both in terms of success about the emergence of language and it relates to the study of important parallel issues, such as human-computer interaction or the very genesis of language.

Challenges for the Multilingual Web of Data

The Web has witnessed an enormous growth in the amount of semantic information published in recent years. This growth has been stimulated to a large extent by the emergence of Linked Data. Although this brings us a big step closer to the vision of a Semantic Web, it also raises new issues such as the need for dealing with information expressed in different natural languages. Indeed, although the Web of Data can contain any kind of information in any language, it still lacks explicit mechanisms to automatically reconcile such information when it is expressed in different languages. This leads to situations in which data expressed in a certain language is not easily accessible to speakers of other languages. The Web of Data shows the potential for being extended to a truly multilingual web as vocabularies and data can be published in a language-independent fashion, while associated language-dependent (linguistic) information supporting the access across languages can be stored separately. In this sense, the multilingual Web of Data can be realized in our view as a layer of services and resources on top of the existing Linked Data infrastructure adding i) linguistic information for data and vocabularies in different languages, ii) mappings between data with labels in different languages, and iii) services to dynamically access and traverse Linked Data across different languages. In this article we present this vision of a multilingual Web of Data. We discuss challenges that need to be addressed to make this vision come true and discuss the role that techniques such as ontology localization, ontology mapping, and cross-lingual ontology-based information access and presentation will play in achieving this. Further, we propose an initial architecture and describe a roadmap that can provide a basis for the implementation of this vision.

Effect of Biodiversity Changes in Disease Risk: Exploring Disease Emergence in a Plant-Virus System

The effect of biodiversity on the ability of parasites to infect their host and cause disease (i.e. disease risk) is a major question in pathology, which is central to understand the emergence of infectious diseases, and to develop strategies for their management. Two hypotheses, which can be considered as extremes of a continuum, relate biodiversity to disease risk: One states that biodiversity is positively correlated with disease risk (Amplification Effect), and the second predicts a negative correlation between biodiversity and disease risk (Dilution Effect). Which of them applies better to different host-parasite systems is still a source of debate, due to limited experimental or empirical data. This is especially the case for viral diseases of plants. To address this subject, we have monitored for three years the prevalence of several viruses, and virus-associated symptoms, in populations of wild pepper (chiltepin) under different levels of human management. For each population, we also measured the habitat species diversity, host plant genetic diversity and host plant density. Results indicate that disease and infection risk increased with the level of human management, which was associated with decreased species diversity and host genetic diversity, and with increased host plant density. Importantly, species diversity of the habitat was the primary predictor of disease risk for wild chiltepin populations. This changed in managed populations where host genetic diversity was the primary predictor. Host density was generally a poorer predictor of disease and infection risk. These results support the dilution effect hypothesis, and underline the relevance of different ecological factors in determining disease/infection risk in host plant populations under different levels of anthropic influence. These results are relevant for managing plant diseases and for establishing conservation policies for endangered plant species.

Policy processes and collective action in water reforms: the cases of Nicaragua and Suriname

Las reformas de agua en países en desarrollo suelen llevarse a cabo junto con cambios institucionales más profundos o, incluso, acompañadas de cambios constitucionales o de régimen político. Por lo tanto, los marcos institucionales adaptados a países gobernados sujetos al estado de derecho pueden no ser apropiados para contextos gobernados fundamentalmente, al menos en sus inicios, por instituciones informales o poco maduras. Esta tesis toma las reformas de agua como punto de partida y pretende contribuir a la literatura mediante una serie de análisis empíricos tanto del ámbito general como del plano individual o sujeto personal de la política del agua. En el ámbito general, el foco se pone en los factores que explican el fallo de la acción colectiva en dos contextos diferentes: 1) en la implementación de la nueva Ley de Aguas de Nicaragua y 2) en el mantenimiento y revitalización de las instituciones de riego en Surinam. En el plano del individuo, la investigación se centra en las decisiones de los usuarios de los recursos y analiza el papel crítico de las variables sociales para la gestión de los recursos comunes. Para ello, el método de investigación utilizado es mixto, combinando el análisis de entrevistas, encuestas y experimentos. En el ámbito general, los resultados muestran que las principales barreras para la implementación de la nueva Ley de Aguas de Nicaragua podrían tener su reflejo en el lenguaje de la Ley y, por tanto, en la forma en la que se definen y configuran las instituciones incluidas en dicha Ley. Así, la investigación demuestra que la implementación de políticas no puede ser estudiada o entendida sin tener en cuenta tanto el diseño de la propia política como el marco socio-ecológico en el que se enmarca. El contexto específico de Nicaragua remarca la importancia de considerar tanto las instituciones formales como informales en los procesos de transición política. A pesar de que las reformas de agua requieren plazos largos para su implementación, el hecho de que exista una diferencia entre las reglas tal cual se definen formalmente y las reglas que operan en la realidad merece una mayor consideración en el diseño de políticas basadas fundamentalmente en instituciones formales. En el ámbito de la conducta individual, el análisis de la acción colectiva ofrece una serie de observaciones empíricas interesantes. En el caso de Nicaragua, los resultados indican que la intensidad de las relaciones sociales, el tipo de agentes dispuestos a proporcionar apoyo social y el nivel de confianza en la comunidad son factores que explican de manera significativa la participación en la comunidad. Sin embargo, el hecho de que la gestión colectiva de riego se produzca, en la mayoría de casos, en torno a lazos familiares sugiere que las variables de capital social críticas se definen en gran medida en la esfera familiar, siendo difícil que se extiendan fuera de estos nexos. El análisis experimental de los resultados de un juego de uso de recurso común y contribución al bien público muestra que las preferencias pro-sociales de los individuos y la heterogeneidad del grupo en términos de composición por sexo son factores que explican significativamente los resultados y las decisiones de apropiación a lo largo del juego. En términos del diseño de las políticas, es fundamental tener en cuenta las dinámicas de participación y uso de los recursos comunes de manera que los niveles de cooperación puedan mantenerse en el largo plazo, lo cual, como se observa en el caso de Surinam, no es siempre posible. Finalmente, el caso de Surinam es un ejemplo ilustrativo de los procesos de acción colectiva en economías en transición. El análisis del fallo de la acción colectiva en Surinam muestra que los procesos políticos vinculados al período colonial y de independencia explican en gran medida la falta de claridad en las reglas operacionales y colectivas que gobiernan la gestión de los sistemas de riego y drenaje. Los resultados empíricos sugieren que a pesar de que la acción colectiva para la provisión de los servicios de riego y drenaje estaba bien establecida bajo el régimen colonial, la auto-organización no prosperó en un contexto dependiente del apoyo externo y regido fundamentalmente por reglas diseñadas al nivel competencial del gobierno central. El sistema socio-ecológico que se desarrolló durante la transición post-colonial favoreció, así, la emergencia de comportamientos oportunistas, y posteriormente la inoperancia de los Water Boards (WBs) creados en la época colonial. En este sentido, cualquier intento por revitalizar los WBs y fomentar el desarrollo de la auto-organización de los usuarios necesitará abordar los problemas relacionados con los patrones demográficos, incluyendo la distribución de la tierra, el diseño de instituciones y la falta de confianza en el gobierno, además de las inversiones típicas en infraestructura y sistemas de información hidrológicos. El liderazgo del gobierno, aportando empuje de arriba-abajo, es, además, otro elemento imprescindible en Surinam. ABSTRACT Water reforms in developing countries take place along deeper institutional and even constitutional. Therefore, institutional frameworks that might result in positive outcomes in countries governed by the rule of law might not fit in contexts governed mainly by informal or immature institutions. This thesis takes water reforms as the starting point and aims to contribute to the literature by presenting several conceptual and empirical analyses at both general and individual levels. At the general national level, the focus is on the factors explaining failure of collective action in two different settings: 1) in the implementation of the new Nicaraguan Water Law and 2) in sustaining and revitalizing irrigation institutions in Suriname. At the individual level, the research focuses on the actions of resource users and analyzes the critical role of social variables for common pool resources management. For this purpose, the research presented in this thesis makes use of a mixed-method approach, combining interviews, surveys and experimental methods. Overall, the results show that major barriers for the implementation of the new Nicaraguan Water Law have its reflection on the language of the Law and, therefore, on the way institutions are defined and configured. In this sense, our study shows that implementation cannot fruitfully be studied and understood without taking into account both the policy design and the social-ecological context in which it is framed. The specific setting of Nicaragua highlights the relevance of considering both formal and informal institutions when promoting policy transitions. Despite the unquestionable fact that water reforms implementation needs long periods of time, there is still a gap between the rules on paper and the rules on the ground that deserves further attention when proposing policy changes on the basis of formal institutions. At the level of the individual agent, the analysis of collective action provides a number of interesting empirical insights. In the case of Nicaragua, I found that the intensity of social networks, the type of agents willing to provide social support and the level of trust in the community are all significant factors in explaining collective action at community level. However, the fact that most collective irrigation relies on family ties suggests that critical social capital variables might be defined within the family sphere and making it difficult to go beyond it. Experimental research combining a common pool resource and a public good game in Nicaragua shows that individuals’ pro-social traits and group heterogeneity in terms of sex composition are significant variables in explaining efficiency outcomes and effort decisions along the game. Thus, with regard to policy design, it is fundamental to consider carefully the dynamics of agents' participation and use of common pool resources, for sustaining cooperation in the long term, which, as seen in the case and Surinam, is not always possible. The case of Suriname provides a rich setting for the analysis of collective action in transition economies. The analysis of decay of collective irrigation in Suriname shows that the lack of clear operational and collective choice rules appear to be rooted in deeper political processes that date back to the colonial period. The empirical findings suggest that despite collective action for the provision of irrigation and drainage services was well established during the colonial period, self-organization did not flourish in a context governed by colonial state-crafted rules and mostly dependent on external support. The social-ecological system developed during the post-colonial transition process favored the emergence of opportunistic behavior. In this respect, any attempt to revitalize WBs and support self-organization will need to tackle the problems derived from demographic patterns, including land allocation, institutions design and government distrust, in addition to the typical investments in both physical infrastructure and hydrological information systems. The leadership role of the government, acting as a top-down trigger, is another essential element in Suriname.