User-centric need-driven affect modeling for spoken conversational agents: design and evaluation

It is easy to get frustrated at spoken conversational agents (SCAs), perhaps because they seem to be callous. By and large, the quality of human-computer interaction is affected due to the inability of the SCAs to recognise and adapt to user emotional state. Now with the mass appeal of artificially-mediated communication, there has been an increasing need for SCAs to be socially and emotionally intelligent, that is, to infer and adapt to their human interlocutors’ emotions on the fly, in order to ascertain an affective, empathetic and naturalistic interaction. An enhanced quality of interaction would reduce users’ frustrations and consequently increase their satisfactions. These reasons have motivated the development of SCAs towards including socio-emotional elements, turning them into affective and socially-sensitive interfaces. One barrier to the creation of such interfaces has been the lack of methods for modelling emotions in a task-independent environment. Most emotion models for spoken dialog systems are task-dependent and thus cannot be used “as-is” in different applications. This Thesis focuses on improving this, in which it concerns computational modeling of emotion, personality and their interrelationship for task-independent autonomous SCAs. The generation of emotion is driven by needs, inspired by human’s motivational systems. The work in this Thesis is organised in three stages, each one with its own contribution. The first stage involved defining, integrating and quantifying the psychological-based motivational and emotional models sourced from. Later these were transformed into a computational model by implementing them into software entities. The computational model was then incorporated and put to test with an existing SCA host, a HiFi-control agent. The second stage concerned automatic prediction of affect, which has been the main challenge towards the greater aim of infusing social intelligence into the HiFi agent. In recent years, studies on affect detection from voice have moved on to using realistic, non-acted data, which is subtler. However, it is more challenging to perceive subtler emotions and this is demonstrated in tasks such as labelling and machine prediction. In this stage, we attempted to address part of this challenge by considering the roles of user satisfaction ratings and conversational/dialog features as the respective target and predictors in discriminating contentment and frustration, two types of emotions that are known to be prevalent within spoken human-computer interaction. The final stage concerned the evaluation of the emotional model through the HiFi agent. A series of user studies with 70 subjects were conducted in a real-time environment, each in a different phase and with its own conditions. All the studies involved the comparisons between the baseline non-modified and the modified agent. The findings have gone some way towards enhancing our understanding of the utility of emotion in spoken dialog systems in several ways; first, an SCA should not express its emotions blindly, albeit positive. Rather, it should adapt its emotions to user states. Second, low performance in an SCA may be compensated by the exploitation of emotion. Third, the expression of emotion through the exploitation of prosody could better improve users’ perceptions of an SCA compared to exploiting emotions through just lexical contents. Taken together, these findings not only support the success of the emotional model, but also provide substantial evidences with respect to the benefits of adding emotion in an SCA, especially in mitigating users’ frustrations and ultimately improving their satisfactions. Resumen Es relativamente fácil experimentar cierta frustración al interaccionar con agentes conversacionales (Spoken Conversational Agents, SCA), a menudo porque parecen ser un poco insensibles. En general, la calidad de la interacción persona-agente se ve en cierto modo afectada por la incapacidad de los SCAs para identificar y adaptarse al estado emocional de sus usuarios. Actualmente, y debido al creciente atractivo e interés de dichos agentes, surge la necesidad de hacer de los SCAs unos seres cada vez más sociales y emocionalmente inteligentes, es decir, con capacidad para inferir y adaptarse a las emociones de sus interlocutores humanos sobre la marcha, de modo que la interacción resulte más afectiva, empática y, en definitiva, natural. Una interacción mejorada en este sentido permitiría reducir la posible frustración de los usuarios y, en consecuencia, mejorar el nivel de satisfacción alcanzado por los mismos. Estos argumentos justifican y motivan el desarrollo de nuevos SCAs con capacidades socio-emocionales, dotados de interfaces afectivas y socialmente sensibles. Una de las barreras para la creación de tales interfaces ha sido la falta de métodos de modelado de emociones en entornos independientes de tarea. La mayoría de los modelos emocionales empleados por los sistemas de diálogo hablado actuales son dependientes de tarea y, por tanto, no pueden utilizarse "tal cual" en diferentes dominios o aplicaciones. Esta tesis se centra precisamente en la mejora de este aspecto, la definición de modelos computacionales de las emociones, la personalidad y su interrelación para SCAs autónomos e independientes de tarea. Inspirada en los sistemas motivacionales humanos en el ámbito de la psicología, la tesis propone un modelo de generación/producción de la emoción basado en necesidades. El trabajo realizado en la presente tesis está organizado en tres etapas diferenciadas, cada una con su propia contribución. La primera etapa incluyó la definición, integración y cuantificación de los modelos motivacionales de partida y de los modelos emocionales derivados a partir de éstos. Posteriormente, dichos modelos emocionales fueron plasmados en un modelo computacional mediante su implementación software. Este modelo computacional fue incorporado y probado en un SCA anfitrión ya existente, un agente con capacidad para controlar un equipo HiFi, de alta fidelidad. La segunda etapa se orientó hacia el reconocimiento automático de la emoción, aspecto que ha constituido el principal desafío en relación al objetivo mayor de infundir inteligencia social en el agente HiFi. En los últimos años, los estudios sobre reconocimiento de emociones a partir de la voz han pasado de emplear datos actuados a usar datos reales en los que la presencia u observación de emociones se produce de una manera mucho más sutil. El reconocimiento de emociones bajo estas condiciones resulta mucho más complicado y esta dificultad se pone de manifiesto en tareas tales como el etiquetado y el aprendizaje automático. En esta etapa, se abordó el problema del reconocimiento de las emociones del usuario a partir de características o métricas derivadas del propio diálogo usuario-agente. Gracias a dichas métricas, empleadas como predictores o indicadores del grado o nivel de satisfacción alcanzado por el usuario, fue posible discriminar entre satisfacción y frustración, las dos emociones prevalentes durante la interacción usuario-agente. La etapa final corresponde fundamentalmente a la evaluación del modelo emocional por medio del agente Hifi. Con ese propósito se llevó a cabo una serie de estudios con usuarios reales, 70 sujetos, interaccionando con diferentes versiones del agente Hifi en tiempo real, cada uno en una fase diferente y con sus propias características o capacidades emocionales. En particular, todos los estudios realizados han profundizado en la comparación entre una versión de referencia del agente no dotada de ningún comportamiento o característica emocional, y una versión del agente modificada convenientemente con el modelo emocional propuesto. Los resultados obtenidos nos han permitido comprender y valorar mejor la utilidad de las emociones en los sistemas de diálogo hablado. Dicha utilidad depende de varios aspectos. En primer lugar, un SCA no debe expresar sus emociones a ciegas o arbitrariamente, incluso aunque éstas sean positivas. Más bien, debe adaptar sus emociones a los diferentes estados de los usuarios. En segundo lugar, un funcionamiento relativamente pobre por parte de un SCA podría compensarse, en cierto modo, dotando al SCA de comportamiento y capacidades emocionales. En tercer lugar, aprovechar la prosodia como vehículo para expresar las emociones, de manera complementaria al empleo de mensajes con un contenido emocional específico tanto desde el punto de vista léxico como semántico, ayuda a mejorar la percepción por parte de los usuarios de un SCA. Tomados en conjunto, los resultados alcanzados no sólo confirman el éxito del modelo emocional, sino xv que constituyen además una evidencia decisiva con respecto a los beneficios de incorporar emociones en un SCA, especialmente en cuanto a reducir el nivel de frustración de los usuarios y, en última instancia, mejorar su satisfacción.

Desing of an Airport Surface Routing Evaluation Tool

There are many studies related with airport surface routing algorithms, based on different approaches and with different evaluation methods and metrics. So, the need of performing a balanced analysis and comparison using a common framework is evident. This paper presents an implementation of an evaluation tool for airport surface routing algorithms. The routing evaluation tool presented here is based in three basic pillars composed by the airport model, the model and generation of traffic and a comprehensive figure of merit function. The paper includes some example evaluations performed over Barajas Airport with representative traffic samples using several simple routing methods.

Self Learning and self evaluation exercises database in basic mechanics: a theoretical-practical approach

In the School of Mines of the Technical University of Madrid (UPM) the first course of different degrees has been implemented and adapted to the European Higher Educational Area (EHEA). In all of the degrees there is a first semester course which gathers all the contents of basic mechanics: from the first kinematics concepts to the rigid solid plane motion Before the Bologna process took place, the authors had established the final assessment of the theoretical contents through open questions of theoretical-practical character In the present work, the elaboration of a wide database containing theoretical-practical questions that students can access on line is presented. The questions are divided in thirteen different questionnaires composed of a number of questions randomly chosen from a certain group in the database. Each group corresponds to a certain learning objective that the student knows. After answering the questionnaire and checking the grade assigned according to the performance of the student, the pupils can see the correct response displayed on the screen and widely explained by the professors. This represents a 10% of the final grade. As the student can access the questionnaires as many times as they want, the main goal is the self-assessment of each learning objective and therefore, getting the students involved in their own learning process so they can decide how much time they need to acquire the required level.

Semantic Interoperability: Perspectives and Technology Evaluation

Interoperability between semantic technologies is a must because they need to be in communication to interchange ontologies and use them in the distributed and open environment of the SemanticWeb. However, such interoperability is not straightforward due to the high heterogeneity in such technologies. This chapter describes the problem of semantic technology interoperability from two different perspectives. First, from a theoretical perspective by presenting an overview of the different factors that affect interoperability and, second, from a practical perspective by reusing evaluation methods and applying them to six current semantic technologies in order to assess their interoperability.

I feel you: the design and evaluation of a domotic affect-sensitive spoken conversational agent

We describe the work on infusion of emotion into a limited-task autonomous spoken conversational agent situated in the domestic environment, using a need-inspired task-independent emotion model (NEMO). In order to demonstrate the generation of affect through the use of the model, we describe the work of integrating it with a natural-language mixed-initiative HiFi-control spoken conversational agent (SCA). NEMO and the host system communicate externally, removing the need for the Dialog Manager to be modified, as is done in most existing dialog systems, in order to be adaptive. The first part of the paper concerns the integration between NEMO and the host agent. The second part summarizes the work on automatic affect prediction, namely, frustration and contentment, from dialog features, a non-conventional source, in the attempt of moving towards a more user-centric approach. The final part reports the evaluation results obtained from a user study, in which both versions of the agent (non-adaptive and emotionally-adaptive) were compared. The results provide substantial evidences with respect to the benefits of adding emotion in a spoken conversational agent, especially in mitigating users' frustrations and, ultimately, improving their satisfaction.

Federating queries in SPARQL 1.1: syntax, semantics and evaluation

Given the sustained growth that we are experiencing in the number of SPARQL endpoints available, the need to be able to send federated SPARQL queries across these has also grown. To address this use case, the W3C SPARQL working group is defining a federation extension for SPARQL 1.1 which allows for combining graph patterns that can be evaluated over several endpoints within a single query. In this paper, we describe the syntax of that extension and formalize its semantics. Additionally, we describe how a query evaluation system can be implemented for that federation extension, describing some static optimization techniques and reusing a query engine used for data-intensive science, so as to deal with large amounts of intermediate and final results. Finally we carry out a series of experiments that show that our optimizations speed up the federated query evaluation process.

Evaluation of reactivity monitoring techniques at the Yalina – Booster subcritical facility

La gestión de los residuos radiactivos de vida larga producidos en los reactores nucleares constituye uno de los principales desafíos de la tecnología nuclear en la actualidad. Una posible opción para su gestión es la transmutación de los nucleidos de vida larga en otros de vida más corta. Los sistemas subcríticos guiados por acelerador (ADS por sus siglas en inglés) son una de las tecnologías en desarrollo para logar este objetivo. Un ADS consiste en un reactor nuclear subcrítico mantenido en un estado estacionario mediante una fuente externa de neutrones guiada por un acelerador de partículas. El interés de estos sistemas radica en su capacidad para ser cargados con combustibles que tengan contenidos de actínidos minoritarios mayores que los reactores críticos convencionales, y de esta manera, incrementar las tasas de trasmutación de estos elementos, que son los principales responsables de la radiotoxicidad a largo plazo de los residuos nucleares. Uno de los puntos clave que han sido identificados para la operación de un ADS a escala industrial es la necesidad de monitorizar continuamente la reactividad del sistema subcrítico durante la operación. Por esta razón, desde los años 1990 se han realizado varios experimentos en conjuntos subcríticos de potencia cero (MUSE, RACE, KUCA, Yalina, GUINEVERE/FREYA) con el fin de validar experimentalmente estas técnicas. En este contexto, la presente tesis se ocupa de la validación de técnicas de monitorización de la reactividad en el conjunto subcrítico Yalina-Booster. Este conjunto pertenece al Joint Institute for Power and Nuclear Research (JIPNR-Sosny) de la Academia Nacional de Ciencias de Bielorrusia. Dentro del proyecto EUROTRANS del 6º Programa Marco de la UE, en el año 2008 se ha realizado una serie de experimentos en esta instalación concernientes a la monitorización de la reactividad bajo la dirección del CIEMAT. Se han realizado dos tipos de experimentos: experimentos con una fuente de neutrones pulsada (PNS) y experimentos con una fuente continua con interrupciones cortas (beam trips). En el caso de los primeros, experimentos con fuente pulsada, existen dos técnicas fundamentales para medir la reactividad, conocidas como la técnica del ratio bajo las áreas de los neutrones inmediatos y retardados (o técnica de Sjöstrand) y la técnica de la constante de decaimiento de los neutrones inmediatos. Sin embargo, varios experimentos han mostrado la necesidad de aplicar técnicas de corrección para tener en cuenta los efectos espaciales y energéticos presentes en un sistema real y obtener valores precisos de la reactividad. En esta tesis, se han investigado estas correcciones mediante simulaciones del sistema con el código de Montecarlo MCNPX. Esta investigación ha servido también para proponer una versión generalizada de estas técnicas donde se buscan relaciones entre la reactividad el sistema y las cantidades medidas a través de simulaciones de Monte Carlo. El segundo tipo de experimentos, experimentos con una fuente continua e interrupciones del haz, es más probable que sea empleado en un ADS industrial. La versión generalizada de las técnicas desarrolladas para los experimentos con fuente pulsada también ha sido aplicada a los resultados de estos experimentos. Además, el trabajo presentado en esta tesis es la primera vez, en mi conocimiento, en que la reactividad de un sistema subcrítico se monitoriza durante la operación con tres técnicas simultáneas: la técnica de la relación entre la corriente y el flujo (current-to-flux), la técnica de desconexión rápida de la fuente (source-jerk) y la técnica del decaimiento de los neutrones inmediatos. Los casos analizados incluyen la variación rápida de la reactividad del sistema (inserción y extracción de las barras de control) y la variación rápida de la fuente de neutrones (interrupción larga del haz y posterior recuperación). ABSTRACT The management of long-lived radioactive wastes produced by nuclear reactors constitutes one of the main challenges of nuclear technology nowadays. A possible option for its management consists in the transmutation of long lived nuclides into shorter lived ones. Accelerator Driven Subcritical Systems (ADS) are one of the technologies in development to achieve this goal. An ADS consists in a subcritical nuclear reactor maintained in a steady state by an external neutron source driven by a particle accelerator. The interest of these systems lays on its capacity to be loaded with fuels having larger contents of minor actinides than conventional critical reactors, and in this way, increasing the transmutation rates of these elements, that are the main responsible of the long-term radiotoxicity of nuclear waste. One of the key points that have been identified for the operation of an industrial-scale ADS is the need of continuously monitoring the reactivity of the subcritical system during operation. For this reason, since the 1990s a number of experiments have been conducted in zero-power subcritical assemblies (MUSE, RACE, KUCA, Yalina, GUINEVERE/FREYA) in order to experimentally validate these techniques. In this context, the present thesis is concerned with the validation of reactivity monitoring techniques at the Yalina-Booster subcritical assembly. This assembly belongs to the Joint Institute for Power and Nuclear Research (JIPNR-Sosny) of the National Academy of Sciences of Belarus. Experiments concerning reactivity monitoring have been performed in this facility under the EUROTRANS project of the 6th EU Framework Program in year 2008 under the direction of CIEMAT. Two types of experiments have been carried out: experiments with a pulsed neutron source (PNS) and experiments with a continuous source with short interruptions (beam trips). For the case of the first ones, PNS experiments, two fundamental techniques exist to measure the reactivity, known as the prompt-to-delayed neutron area-ratio technique (or Sjöstrand technique) and the prompt neutron decay constant technique. However, previous experiments have shown the need to apply correction techniques to take into account the spatial and energy effects present in a real system and thus obtain accurate values for the reactivity. In this thesis, these corrections have been investigated through simulations of the system with the Monte Carlo code MCNPX. This research has also served to propose a generalized version of these techniques where relationships between the reactivity of the system and the measured quantities are obtained through Monte Carlo simulations. The second type of experiments, with a continuous source with beam trips, is more likely to be employed in an industrial ADS. The generalized version of the techniques developed for the PNS experiments has also been applied to the result of these experiments. Furthermore, the work presented in this thesis is the first time, to my knowledge, that the reactivity of a subcritical system has been monitored during operation simultaneously with three different techniques: the current-to-flux, the source-jerk and the prompt neutron decay techniques. The cases analyzed include the fast variation of the system reactivity (insertion and extraction of a control rod) and the fast variation of the neutron source (long beam interruption and subsequent recovery).

Seismic performance and damage evaluation of a reinforced concrete frame with hysteretic dampers through shake-table tests

Passive energy dissipation devices are increasingly implemented in frame structures to improve their performance under seismic loading. Most guidelines for designing this type of system retain the requirements applicable to frames without dampers, and this hinders taking full advantage of the benefits of implementing dampers. Further, assessing the extent of damage suffered by the frame and by the dampers for different levels of seismic hazard is of paramount importance in the framework of performance-based design. This paper presents an experimental investigation whose objectives are to provide empirical data on the response of reinforced concrete (RC) frames equipped with hysteretic dampers (dynamic response and damage) and to evaluate the need for the frame to form a strong column-weak beam mechanism and dissipate large amounts of plastic strain energy. To this end, shake-table tests were conducted on a 2/5-scale RC frame with hysteretic dampers. The frame was designed only for gravitational loads. The dampers provided lateral strength and stiffness, respectively, three and 12 times greater than those of the frame. The test structure was subjected to a sequence of seismic simulations that represented different levels of seismic hazard. The RC frame showed a performance level of "immediate occupancy", with maximum rotation demands below 20% of the ultimate capacity. The dampers dissipated most of the energy input by the earthquake. It is shown that combining hysteretic dampers with flexible reinforced concrete frames leads to structures with improved seismic performance and that requirements of conventional RC frames (without dampers) can be relieved.

Towards a Learning Object pedagogical quality metric based on the LORI evaluation model

Evaluating and measuring the pedagogical quality of Learning Objects is essential for achieving a successful web-based education. On one hand, teachers need some assurance of quality of the teaching resources before making them part of the curriculum. On the other hand, Learning Object Repositories need to include quality information into the ranking metrics used by the search engines in order to save users time when searching. For these reasons, several models such as LORI (Learning Object Review Instrument) have been proposed to evaluate Learning Object quality from a pedagogical perspective. However, no much effort has been put in defining and evaluating quality metrics based on those models. This paper proposes and evaluates a set of pedagogical quality metrics based on LORI. The work exposed in this paper shows that these metrics can be effectively and reliably used to provide quality-based sorting of search results. Besides, it strongly evidences that the evaluation of Learning Objects from a pedagogical perspective can notably enhance Learning Object search if suitable evaluations models and quality metrics are used. An evaluation of the LORI model is also described. Finally, all the presented metrics are compared and a discussion on their weaknesses and strengths is provided.

Method for electrical evaluation of graphene using a GFET structure

n this work, we explain a method to characterize graphene using electrical measurements in graphene field-effect transistors (GFET) devices. Our goal is to obtain the material electronic properties from the output characteristics of one GFET device. For the previous purpose, we will need to apply a physical model that allows us to correlate the electronic behavior of a GFET with the material properties.

SemQuaRE - An extension of the SQuaRE quality model for the evaluation of semantic technologies

To correctly evaluate semantic technologies and to obtain results that can be easily integrated, we need to put evaluations under the scope of a unique software quality model. This paper presents SemQuaRE, a quality model for semantic technologies. SemQuaRE is based on the SQuaRE standard and describes a set of quality characteristics specific to semantic technologies and the quality measures that can be used for their measurement. It also provides detailed formulas for the calculation of such measures. The paper shows that SemQuaRE is complete with respect to current evaluation trends and that it has been successfully applied in practice.