This paper researches the theoretical background needed for the implementation of literary texts with multicultural themes for use in EFL courses in Sweden and it offers several concrete didactical solutions for multicultural reading. The theory of multicultural education by J. A. Banks is presented with focus on the dimension of knowledge construction. The processes that are examined are the learning processes in the zone of proximal development by Vygotsky (1986), the concept of scaffolding by Woods, Bruner & Ross (1976) and the process of perspective-taking by Thein & Sloan (2013). These processes each employ a three-step sequence that moves students from their existing knowledge to new knowledge and revised personal opinions. The teacher’s role is to provide support during the learning process. The second part of the paper suggests different activities for the multicultural reading of a novel, used to make the process of multicultural knowledge construction and scaffolding visible. Teacher support includes text reduction, book discussion and language analysis. Discussion points found in Love Medicine start with revising the stereotypical images, discovering how personal experience influences knowledge or how ethnicity influences professional career choices. When teachers and students read literary texts with multicultural themes, students’ racial prejudice can be reduced (Banks 2004) and their ethical attitudes become more open (Thein & Sloan 2013). The literary work chosen for framing in the theory is Love Medicine by Louise Erdrich. The paper ends with a reflection over the limitations of multicultural reading projects.

Un système de dialogue permet de doter la Machine de la capacité d'interagir de façon naturelle et efficace avec l'Homme. Dans cette thèse nous nous intéressons au développement d'un système de dialogue reposant sur des approches statistiques, et en particulier du cadre formel des Processus Décisionnel de Markov Partiellement Observable, en anglais Partially Observable Markov Decision Process (POMDP), qui à ce jour fait office de référence dans la littérature en ce qui concerne la gestion statistique du dialogue. Ce modèle permet à la fois une prise en compte améliorée de l'incertitude inhérente au traitement des données en provenance de l'utilisateur (notamment la parole) et aussi l'optimisation automatique de la politique d'interaction à partir de données grâce à l'apprentissage par renforcement, en anglais Reinforcement Learning (RL). Cependant, une des problématiques liées aux approches statistiques est qu'elles nécessitent le recours à une grande quantité de données d'apprentissage pour atteindre des niveaux de performances acceptables. Or, la collecte de telles données est un processus long et coûteux qui nécessite généralement, pour le cas du dialogue, la réalisation de prototypes fonctionnels avec l'intervention d'experts et/ou le développement de solution alternative comme le recours à la simulation d'utilisateurs. En effet, très peu de travaux considèrent à ce jour la possibilité d'un apprentissage de la stratégie de la Machine de part sa mise en situation de zéro (sans apprentissage préalable) face à de vrais utilisateurs. Pourtant cette solution présente un grand intérêt, elle permet par exemple d'inscrire le processus d'apprentissage comme une partie intégrante du cycle de vie d'un système lui offrant la capacité de s'adapter à de nouvelles conditions de façon dynamique et continue. Dans cette thèse, nous nous attacherons donc à apporter des solutions visant à rendre possible ce démarrage à froid du système mais aussi, à améliorer sa capacité à s'adapter à de nouvelles conditions (extension de domaine, changement d'utilisateur,...). Pour ce faire, nous envisagerons dans un premier temps l'utilisation de l'expertise du domaine (règles expertes) pour guider l'apprentissage initial de la politique d'interaction du système. De même, nous étudierons l'impact de la prise en compte de jugements subjectifs émis par l'utilisateur au fil de l'interaction dans l'apprentissage, notamment dans un contexte de changement de profil d'utilisateur où la politique préalablement apprise doit alors pouvoir s'adapter à de nouvelles conditions. Les résultats obtenus sur une tâche de référence montrent la possibilité d'apprendre une politique (quasi-)optimale en quelques centaines d'interactions, mais aussi que les informations supplémentaires considérées dans nos propositions sont à même d'accélérer significativement l'apprentissage et d'améliorer la tolérance aux bruits dans la chaîne de traitement. Dans un second temps nous nous intéresserons à réduire les coûts de développement d'un module de compréhension de la parole utilisé dans l'étiquetage sémantique d'un tour de dialogue. Pour cela, nous exploiterons les récentes avancées dans les techniques de projection des mots dans des espaces vectoriels continus conservant les propriétés syntactiques et sémantiques, pour généraliser à partir des connaissances initiales limitées de la tâche pour comprendre l'utilisateur. Nous nous attacherons aussi à proposer des solutions afin d'enrichir dynamiquement cette connaissance et étudier le rapport de cette technique avec les méthodes statistiques état de l'art. Là encore nos résultats expérimentaux montrent qu'il est possible d'atteindre des performances état de l'art avec très peu de données et de raffiner ces modèles ensuite avec des retours utilisateurs dont le coût peut lui-même être optimisé. / A dialogue system should give the machine the ability to interactnaturally and efficiently with humans. In this thesis, we focus on theissue of the development of stochastic dialogue systems. Thus, we especiallyconsider the Partially Observable Markov Decision Process (POMDP)framework which yields state-of-the-art performance on goal-oriented dialoguemanagement tasks. This model enables the system to cope with thecommunication ambiguities due to noisy channel and also to optimize itsdialogue management strategy directly from data with Reinforcement Learning (RL)methods.Considering statistical approaches often requires the availability of alarge amount of training data to reach good performance. However, corpora of interest are seldom readily available and collectingsuch data is both time consuming and expensive. For instance, it mayrequire a working prototype to initiate preliminary experiments with thesupport of expert users or to consider other alternatives such as usersimulation techniques.Very few studies to date have considered learning a dialogue strategyfrom scratch by interacting with real users, yet this solution is ofgreat interest. Indeed, considering the learning process as part of thelife cycle of a system offers a principle framework to dynamically adaptthe system to new conditions in an online and seamless fashion.In this thesis, we endeavour to provide solutions to make possible thisdialogue system cold start (nearly from scratch) but also to improve its ability to adapt to new conditions in operation (domain extension, new user profile, etc.).First, we investigate the conditions under which initial expertknowledge (such as expert rules) can be used to accelerate the policyoptimization of a learning agent. Similarly, we study how polarized userappraisals gathered throughout the course of the interaction can beintegrated into a reinforcement learning-based dialogue manager. Morespecifically, we discuss how this information can be cast intosocially-inspired rewards to speed up the policy optimisation for bothefficient task completion and user adaptation in an online learning setting.The results obtained on a reference task demonstrate that a(quasi-)optimal policy can be learnt in just a few hundred dialogues,but also that the considered additional information is able tosignificantly accelerate the learning as well as improving the noise tolerance.Second, we focus on reducing the development cost of the spoken language understanding module. For this, we exploit recent word embedding models(projection of words in a continuous vector space representing syntacticand semantic properties) to generalize from a limited initial knowledgeabout the dialogue task to enable the machine to instantly understandthe user utterances. We also propose to dynamically enrich thisknowledge with both active learning techniques and state-of-the-artstatistical methods. Our experimental results show that state-of-the-artperformance can be obtained with a very limited amount of in-domain andin-context data. We also show that we are able to refine the proposedmodel by exploiting user returns about the system outputs as well as tooptimize our adaptive learning with an adversarial bandit algorithm tosuccessfully balance the trade-off between user effort and moduleperformance.Finally, we study how the physical embodiment of a dialogue system in a humanoid robot can help the interaction in a dedicated Human-Robotapplication where dialogue system learning and testing are carried outwith real users. Indeed, in this thesis we propose an extension of thepreviously considered decision-making techniques to be able to take intoaccount the robot's awareness of the users' belief (perspective taking)in a RL-based situated dialogue management optimisation procedure.

Let’s imagine a typical word-learning scenario. A toddler sits in her highchair in the kitchen and waits for her lunch. Her mother says: “Use a spoon to eat your meal”. Several objects are placed in front of the child. She can see a dish with her lunch, a spoon, a cup, a sugar bowl, a milk jar, her mother’s plate and a second cup. All these objects, present in the visual array, must be identified by the toddler; she must also parse the auditory stream into segments and determine which words are familiar and which ones are potentially new. If the child does not know the word “spoon”, she will need to use the event of naming of this referent by her mother to adjust her attention to the relevant referent. She also needs to update her representation of this word upon hearing it in different contexts with different speakers and, perhaps, different types of spoons. Efficient attention allocation in this word-learning situation will clearly contribute to the success of mapping; the degree of encoding of the word-form and of its meaning will certainly influence whether this word enters the child’s vocabulary.The complexity of such a typical scenario seems very challenging for a toddler whose cognitive resources are still far from being fully mature. Unsurprisingly, several accounts of how toddlers manage to solve this task are currently on the market. The problem of ambiguity associated with meaning-to-referent mapping (several objects co-presented in the visual scene) and with word form-to-meaning mapping (the correct word is to be singled out among phonological competitors) might be even more challenging for children who present an atypical developmental trajectory.Children with Autism Spectrum Disorder and toddlers at risk for the disorder were found to acquire their vocabularies at a slower rate than their typically developing peers. In the contemporary literature, this delay in lexical knowledge acquisition is associated with poor socio-pragmatic understanding that presumably limits children’s capacity to establish referents for words in social contexts. Since impaired understanding of social interaction is a core characteristic of the cognitive profile of individuals with ASD, such an explanation of the delay in language development seems very plausible. However, several other theoretical accounts hold that in typical development socio-pragmatic skills emerge, bottom-up, through more domain-general processing of interactional experiences. In line with the latter views, it can be hypothesized that delays in lexical acquisition in ASD are not directly linked to poor socio-pragmatic understanding but are caused by low-level deficits and atypical attention allocation during word learning.Research programs on lexical learning and processing in ASD thus face the existence of different, contradictory theories of first language acquisition in typical development. Deciding a priori to build one’s experimental study against this or that theoretical background carries the risk of a limited interpretation of experimental results. A more promising way to deal with the variety of available theories of language acquisition is rather to directly confront the existing paradigms and to plan the study design in accordance. This is the approach that I privilege here. In the studies presented within this thesis, I question how social cues are used to resolve ambiguity in meaning during word-learning tasks (chapter 1) and during referential processing in typically developing children (chapter 2) and in children with ASD (chapter 3). Not only do I attempt to compare the use of social cues in word-learning and of perspectival information in referential processing in children with and without ASD, but I also try to link these results with two opposing theoretical views: the one that postulates early reliance on socio-pragmatic understanding and the other that conceives of word-learning as not being necessarily grounded in social understanding. In Chapter 1, I present evidence that children with ASD, children with SLI and typically developing children learn novel words in a flexible way by selectively attending to mappings offered by previously accurate speakers. However, I also show that such learning is likely to be supported by a surface trait attribution mechanism, rather than by genuine socio-pragmatic understanding: children in both clinical groups fail to learn selectively, when learning requires genuinely building a model of the speakers’ epistemic states. Chapters 2 and 3 are devoted to referential communication. I adopt several analytical and methodological modifications to existing methods, which allows me to compare two different aspects of partner-dependent processing of referential precedents. Typically developing children can be expected to recognize precedents previously established with the same partner faster, because of an automatic priming mechanism. However, potential faster processing of broken precedents with a new partner could not be explained by a low-level memory mechanism and would strongly suggest that lexical processing is influenced by expectations about the child’s partner perspective. I present evidence that children with and without ASD do not spontaneously rely on common ground during referential processing and that partner-specific effects in processing are associated with low-level priming. In chapter 3, I report evidence of impaired ability to switch between different conceptual perspectives in children with ASD, which may lead to maladaptive behavior in communication. In the last chapter of this thesis, I explore how word form-to-meaning ambiguity is resolved in children with ASD and whether these children exhibit difficulty in correctly mapping similar-sounding novel words. The results of this study suggest that lexical activation in children with ASD may be impaired and they display deficits in suppressing phonological competitors. Taken together, the results presented in this doctoral dissertation suggest that delays in word acquisition in ASD are likely to be driven by deficits in domain-general cognitive development. Even though impaired socio-cognitive understanding may lead to difficulties in discourse and pragmatics in older children, delayed access to lexical meaning in young children with ASD is likely to be associated with disruptions in domain-general mechanisms of perception, attention and memory. / Doctorat en Langues, lettres et traductologie / info:eu-repo/semantics/nonPublished

La prise de perspective (PP) renvoie à la capacité de comprendre le point de vue psychologique d’autrui. Elle constitue une notion complexe, impliquant différents processus, comme la capacité de prendre conscience de soi et d’autrui, d’inhiber son point de vue, de passer d’un point de vue à un autre ou d’endosser la perspective d’autrui. Ces étapes mobilisent la conscience de soi (CS), la distinction soi-autrui (DSA) et la flexibilité cognitive (FC).Notre recherche vise à évaluer comment des déficits dans chacune de ces sphères pourraient sous-tendre une altération des capacités de PP. Pour ce faire, trois protocoles originaux sont utilisés : ils s’attachent à clarifier les liens entre la PP et les autres composantes empathiques à l’aide de mesures physiologiques, à évaluer les déficits des composantes évoquées dans la dépression sous-clinique et à tester l’effet de deux programmes d’entraînements (CF, PP) sur la symptomatologie dépressive.Les résultats montrent une implication de la PP comme processus transversal des réponses empathiques et son lien avec les composantes automatiques de l’empathie. Ils soutiennent l’existence d’un pattern de déficits en CS, DSA, FC et PP semblables à ceux mis en évidence dans les niveaux cliniques de dépression. Enfin, l’efficacité d’un entraînement à la CF sur les capacités de PP ajoute un argument au lien causal unissant les deux concepts. Nos résultats contribuent à mieux comprendre la dynamique commandant à une chaîne de déficits inter-reliés pour mieux prévenir la maladie chez les personnes à risques. / Perspective-taking (PT) refers to the ability to understand the psychological point of view of others. It is a complex notion involving the ability to become aware of oneself and others, to inhibit one's point of view, to move from one point of view to another and to adopt the perspective of others. These various stages mobilize self-consciousness (CS), self-other distinction (SOD) and cognitive flexibility (CF).Our research aims to evaluate how deficits in each of these spheres could underlie an alteration of perspective-taking abilities. To do so, it uses three original protocols aimed at clarifying the links between PT and the other empathic components by using physiological measures, evaluating the deficits of the processes mentioned in subclinical depression and to test the effect of two training programs (CF, PT) on depressive symptomatology.The results show PT involvement as a transversal process of empathic responses and its connection to the automatic components of empathy. They also support the existence of a pattern of deficits in CS, SOD, FC and PT similar to those found in clinical depression levels. Finally, the effectiveness of PT training on PT capabilities adds an argument to the causal link between these two concepts. Our results contribute to a better understanding of the dynamics driving a chain of interrelated deficits to better prevent the emergence of depression among people at risk.

Unresolved global problems, such as extreme poverty, ask for a better understanding of what predicts altruism and what does not. The aim of this thesis project was to address this topical and timely research question by studying the predictive role of compassion and empathy in understanding altruism. In past research on the relationship between altruism and empathy, distinct empathic processes (Perspective taking, Empathic concern, Personal distress, Emotional contagion, and Behavioral contagion) have been often lumped together and the context dependency of the relationship has been insufficiently taken into account, resulting in confusion and contradictory findings. Compassion overcomes these issues. The present web-based survey with previous or current university students (age 18-45; N=240) aimed to clarify relationships between components of empathy, compassion, and altruism. It was hypothesized that (1) compassion would predict altruism beyond all components of empathy; (2) Empathic concern would mediate the relationship between Perspective taking and altruism; (3) compassion would mediate the relationship between Empathic concern and altruism, and (4) higher levels of compassion would result in a reduced negative relationship between Personal distress and altruism. The results supported all hypotheses except for the final one. These findings are discussed in context of previous research and theory, considering the current study limitations and with focus on theoretical and practical implications. In sum, the findings suggest that efforts to motivate altruism should focus on invoking positive emotions of warmth, concern, and relatability. Care should be taken to avoid unnecessary Personal distress when invoking altruism, as this reduces its likelihood.

No description available.

Les premiers robots sont apparus dans les usines, sous la forme d'automates programmables. Ces premières formes robotiques ont le plus souvent un nombre très limité de capteurs et se contentent de répéter une séquence de mouvements et d'actions. De nos jours, de plus en plus de robots ont à interagir ou coopérer avec l'homme, que se soit sur le lieu de travail avec les robots coéquipiers ou dans les foyers avec les robots d'assistance. Mettre un robot dans un environnement humain soulève de nombreuses problématiques. En effet, pour évoluer dans le même environnement que l'homme et comprendre cet environnement, le robot doit être doté de capacités cognitives appropriées. Au delà de la compréhension de l'environnement matériel, le robot doit être capable de raisonner sur partenaires humains afin de pouvoir collaborer avec eux ou les servir au mieux. Lorsque le robot interagit avec des humains, l'accomplissement de la tâche n'est pas un critère suffisant pour quantifier la qualité de l'interaction. En effet, l'homme étant un être social, il est important que le robot puisse avoir des mécanismes de raisonnement lui permettant d'estimer également l'état mental de l'homme pour améliorer sa compréhension et son efficacité, mais aussi pour exhiber des comportements sociaux afin de se faire accepter et d'assurer le confort de l'humain. Dans ce manuscrit, nous présentons tout d'abord une infrastructure logicielle générique (indépendante de la plateforme robotique et des capteurs utilisés) qui permet de construire et maintenir une représentation de l'état du monde à l'aide de l'agrégation des données d'entrée et d'hypothèses sur l'environnement. Cette infrastructure est également en charge de l'évaluation de la situation. En utilisant l'état du monde qu'il maintient à jour, le système est capable de mettre en oeuvre divers raisonnements spatio-temporels afin d'évaluer la situation de l'environnement et des agents (humains et robots) présents. Cela permet ainsi d'élaborer et de maintenir une représentation symbolique de l'état du monde et d'avoir une connaissance en permanence de la situation des agents. Dans un second temps, pour aller plus loin dans la compréhension de la situation des humains, nous expliquerons comment nous avons doté notre robot de la capacité connue en psychologie développementale et cognitive sous le nom de “théorie de l'esprit” concrétisée ici par des mécanismes permettant de raisonner en se mettant à la place de l'humain, c'est à dire d'être doté de “prise de perspective”. Par la suite nous expliquerons comment l'évaluation de la situation permet d'établir un dialogue situé avec l'homme, et en quoi la capacité de gérer explicitement des croyances divergentes permet d'améliorer la qualité de l'interaction et la compréhension de l'homme par le robot. Nous montrerons également comment la connaissance de la situation et la possibilité de raisonner en se mettant à la place de l'homme permet une reconnaissance d'intentions appropriée de celui-ci et comment nous avons pu grâce à cela doter notre robot de comportements proactifs pour venir en aide à l'homme . Pour finir, nous présenterons une étude présentant un système de maintien d'un modèle des connaissances de l'homme sur diverses tâches et qui permet une gestion adaptée de l'interaction lors de l'élaboration interactive et l'accomplissement d'un plan partagé. / The first robots appeared in factories, in the form of programmable controllers. These first robotic forms usually had a very limited number of sensors and simply repeated a small set of sequences of motions and actions. Nowadays, more and more robots have to interact or cooperate with humans, whether at the workplace with teammate robots or at home with assistance robots. Introducing a robot in a human environment raises many challenges. Indeed, to evolve in the same environment as humans, and to understand this environment, the robot must be equipped with appropriate cognitive abilities. Beyond understanding the physical environment, the robot must be able to reason about human partners in order to work with them or serve them best. When the robot interacts with humans, the fulfillment of the task is not a sufficient criterion to quantify the quality of the interaction. Indeed, as the human is a social being, it is important that the robot can have reasoning mechanisms allowing it to assess the mental state of the human to improve his understanding and efficiency, but also to exhibit social behaviors in order to be accepted and to ensure the comfort of the human. In this manuscript, we first present a generic framework (independent of the robotic platform and sensors used) to build and maintain a representation of the state of the world by using the aggregation of data entry and hypotheses on the environment. This infrastructure is also in charge of assessing the situation. Using the state of the world it maintains, the system is able to utilize various spatio-temporal reasoning to assess the situation of the environment and the situation of the present agents (humans and robots). This allows the creation and maintenance of a symbolic representation of the state of the world and to keep awareness of each agent status. Second, to go further in understanding the situation of the humans, we will explain how we designed our robot with the capacity known in developmental and cognitive psychology as "theory of mind", embodied here by mechanisms allowing the system to reason by putting itself in the human situation, that is to be equipped with "perspective-taking" ability. Later we will explain how the assessment of the situation enables a situated dialogue with the human, and how the ability to explicitly manage conflicting beliefs can improve the quality of interaction and understanding of the human by the robot. We will also show how knowledge of the situation and the perspective taking ability allows proper recognition of human intentions and how we enhanced the robot with proactive behaviors to help the human. Finally, we present a study where a system maintains a human model of knowledge on various tasks to improve the management of the interaction during the interactive development and fulfillment of a shared plan.

This paper discusses the design, implementation, and evaluation of a serious game intended to reinforce applied behavior analysis (ABA) techniques used with children with autism spectrum disorder (ASD) by providing a low cost and easily accessible supplement to traditional methods. Past and recent research strongly supports the use of computer assisted instruction in the education of individuals with ASD (Moore & Calvert, 2000; Noor, Shahbodin, & Pee, 2012). Computer games have been shown to boost confidence and provide calming mechanisms (Griffiths, 2003) while being a safe environment for social exploration and learning (Moore, Cheng, McGrath, & Powell, 2005). Games increase children's motivation and thus increase the rate of learning in computer mediated environments (Moore & Calvert, 2000). Furthermore, children with ASD are able to understand basic emotions and facial expressions in avatars more easily than in real-world interactions (Moore, Cheng, McGrath, & Powell, 2005). Perspective-taking (also known as role-taking) has been shown to be a crucial component and antecedent to empathy (Gomez-Becerra, Martin, Chavez-Brown, & Greer, 2007; Peng, Lee, & Heeter, 2010). Though symptoms vary across children with ASD, perspective-taking and empathy are abilities that have been shown to be limited across a wide spectrum of individuals with ASD and Asperger's disorder (Gomez-Becerra, Martin, Chavez-Brown, & Greer, 2007). A game called WUBeeS was developed to aid young children with ASD in perspective taking and empathy by placing the player in the role of a caregiver to a virtual avatar. It is hypothesized that through the playing of this game over a series of trials, children with ASD will show an iv increase in the ability to discriminate emotions, provide appropriate responses to basic needs (e.g. feeding the avatar when it is hungry), and be able to communicate more clearly about emotions.

This thesis contributes to our understanding of cognitive interoperability by explor-ing barriers, facilitators, and contextual factors to create a framework. With the ad-vent of the cognitive domain in warfighting, the adversary pursuit of military strate-gic advantage through cognitive science demands that we seize the initiative and seek cognitive superiority with allies and partners. Prior research acknowledges the importance of human interoperability but is limited to singular studies at the opera-tional and national level. This thesis shifts focus to the military strategic and multi-national level to uncover cognitive and cultural inhibitors and enablers of cognitive interoperability. The empirical data is drawn from a distinct case study that examines senior military officers during a combined exercise. The analysis explores compe-tencies that foster strategic empathy and collective intentionality with identity and human connectivity as major catalysts. Individuals are agents who collectively con-struct cognitive interoperability, setting conditions for cognitive dominance in future military competition.

Based on a targeted literature review, this vision paper emphasizes the importance of dementia-sensitive built space. The article specifically focuses on supporting spatial orientation and wayfinding for people living with dementia. First, we discuss types of wayfinding challenges, underlying processes, and consequences of spatial disorientation in the context of dementia of the Alzheimer’s type. Second, we focus on current efforts aimed at planning and evaluating dementia-sensitive built space, i.e., environmental design principles, interventions, evaluation tools, strategies, and planning processes. Third, we use our findings as a starting point for developing an interdisciplinary research vision aimed at encouraging further debates and research about: (1) the perspective of a person with dementia, specifically in the context of wayfinding and spatial orientation, and (2) how this perspective supplements planning and design processes of dementia-sensitive built space. We conclude that more closely considering the perspective of people with dementia supports the development of demographically sustainable future cities and care institutions.