Many companies implemented remote work procedures during the pandemic and for many organizations video meetings have since remained a staple. Remote working has enabled employees to be more flexible with their schedules and technical solutions such as live captioning has been identified as potentially enabling deaf/hard-of-hearing employees during meetings. However with new procedures comes concern about how we potentially can be affected by the changes. Some earlier research has shown that speech intelligibility can be improved by the inclusion of text support, but they also raised the possibility that it could have unwanted adverse effects on cognitive abilities (Zhong, Noud et al., 2022). This study was conducted with this focus, studying the effects of text support on specifically listening effort and comprehension during normal as well as adverse conditions (featuring added noise). To investigate the effects of text support a 2 (Noise, No Noise) x 2 (Text Support, No Text Support) design was used. The participants were shown 16 short videos simulating video meetings and after each video were asked to rate their perceived listening effort as well as a comprehension question about the contents of the discussion. Each of the four conditions were equally represented but the order of the specific video files and conditions that applied were randomised for each participant to mitigate undue effects. The results of the study indicate that the presence of captions decrease effort and raise comprehension in both normal and adverse conditions. Noise was found to strongly effect the listening effort required by participants but no significant effect was found upon comprehension. Some concerns regarding the ecological validity were identified during the course of the study such as only studying energetic noise and unrealistic presentation of captions. However the results are nonetheless believed to be generalizable in most regards and showcase that captions can have a positive influence during video meetings.

No description available.

Universally across modern cultures children acquire the meaning of the words one, two, and three in order. While much research has focused on how children acquire this knowledge and what this knowledge represents, the question of why children learn numbers in order has been comparatively neglected. To address this question, a non-verbal anticipatory looking task was implemented. In this task, 35 14- to 23-month-old infants were assessed on their ability to form implicit category structures for the numbers one, two, and three. We hypothesized that children would be able to form the implicit category structure for the number one but not for two or three because sets of two and three objects would exceed the working memory capacities of infants. We found results consistent with this hypothesis; infants (regardless of age) were able form a category for sets with one object, as evidenced by their looking behavior while the looking behavior for the numbers two and three did not demonstrate a statistically significant pattern. We interpret our results as consistent with our hypothesis and discuss implications for parallel individuation, number acquisition theories, and the development of working memory resources.

Social attitude is the approach of a person displayed towards other individuals or groups. Social attitude comprehensively affects the way we perceive, behave in, and interact with, the surrounding world; it is simply not possible to understand complex social behaviour such as strategic thinking without first knowing the attitude of the parties involved. Several disciplines contribute to the complex study of social attitude (social preferences in economics, social value orientation in psychology), but only recently have these disciplines started to communicate and develop comprehensive definitions and models. In particular, the current research debate focuses on pinpointing the nature of social attitude (e.g., what its defining components are), the factors that influence it (e.g. context, other individuals), as well as its consequences (e.g., its relevance for self-image representation). This thesis aims to answer to some of the open questions in the literature by testing and comparing the proposed competing explanations. The studies presented are based on a series of behavioural experiments coupled with established but also newly developed measurement tools concerning social norms and personal preferences. In addition, we try to uncover the mental processes underlying decisions with the help of computational models. The thesis is structured as follows. In Chapter 1, We outline a brief summary of the theories on social attitude from the economic and psychological literature, and describe the main tasks and models employed in the thesis. Chapter 2 explores how social attitude is influenced by others’ behaviour. We conduct a systematic comparison of the possible mechanisms driving attitude conformity using various experimental conditions, computational models, and control tasks (e.g., norm elicitation). We find that participants conform due to both peer influence (by learning from others about how salient a norm is) and compliance to authority (i.e. experimenter demand effects). Chapter 3 studies the effect of context in a task eliciting social attitude. We specifically test the effect of unavailable choices, that we call ”meta-context”, on participant’s decisions. We find that participants’ concerns about social norms, as well as their choices, depend on the currently available options, but also on meta-context. In Chapter 4, we study whether individuals tend to selectively forget about their morally questionable choices, and information related to it, such as the context in which the choice was made. We find that participants recollect less correctly selfish or anti-social choices compared to pro-social ones, but we find no memory bias concerning the context of the choice. Moreover, we uncover some potential evidence of a second memory bias related to choice frequency: people are generally more pro-social than antisocial, which means antisocial choices are more rare and thus more difficult to remember correctly. Finally, in chapter 5 We summarise the main findings of the thesis and present some conclusions. We try to integrate the various results to propose an empirically-informed model of social attitude to be applied in future research on the topic.

This work proposes a strategy for visual perception in the context of autonomous driving. Despite the growing research aiming to implement self-driving cars, no artificial system can claim to have reached the driving performance of a human, yet. Humans---when not distracted or drunk---are still the best drivers you can currently find. Hence, the theories about the human mind and its neural organization could reveal precious insights on how to design a better autonomous driving agent. This dissertation focuses specifically on the perceptual aspect of driving, and it takes inspiration from four key theories on how the human brain achieves the cognitive capabilities required by the activity of driving. The first idea lies at the foundation of current cognitive science, and it argues that thinking nearly always involves some sort of mental simulation, which takes the form of imagery when dealing with visual perception. The second theory explains how the perceptual simulation takes place in neural circuits called convergence-divergence zones, which expand and compress information to extract abstract concepts from visual experience and code them into compact representations. The third theory highlights that perception---when specialized for a complex task as driving---is refined by experience in a process called perceptual learning. The fourth theory, namely the free-energy principle of predictive brains, corroborates the role of visual imagination as a fundamental mechanism of inference. In order to implement these theoretical principles, it is necessary to identify the most appropriate computational tools currently available. Within the consolidated and successful field of deep learning, I select the artificial architectures and strategies that manifest a sounding resemblance with their cognitive counterparts. Specifically, convolutional autoencoders have a strong correspondence with the architecture of convergence-divergence zones and the process of perceptual abstraction. The free-energy principle of predictive brains is related to variational Bayesian inference and the use of recurrent neural networks. In fact, this principle can be translated into a training procedure that learns abstract representations predisposed to predicting how the current road scenario will change in the future. The main contribution of this dissertation is a method to learn conceptual representations of the driving scenario from visual information. This approach forces a semantic internal organization, in the sense that distinct parts of the representation are explicitly associated to specific concepts useful in the context of driving. Specifically, the model uses as few as 16 neurons for each of the two basic concepts here considered: vehicles and lanes. At the same time, the approach biases the internal representations towards the ability to predict the dynamics of objects in the scene. This property of temporal coherence allows the representations to be exploited to predict plausible future scenarios and to perform a simplified form of mental imagery. In addition, this work includes a proposal to tackle the problem of opaqueness affecting deep neural networks. I present a method that aims to mitigate this issue, in the context of longitudinal control for automated vehicles. A further contribution of this dissertation experiments with higher-level spaces of prediction, such as occupancy grids, which could conciliate between the direct application to motor controls and the biological plausibility.

Embodied Conversational Agents (ECA) form part of a range of virtual characters whose intended purpose include engaging in natural conversations with human users. While works in literature are ripe with descriptions of attempts at producing viable ECA architectures, few authors have addressed the role of episodic memory models in conversational agents. This form of memory, which provides a sense of autobiographic record-keeping in humans, has only recently been peripherally integrated into dialog management tools for ECAs. In our work, we propose to take a closer look at the shared characteristics of episodic memory models in recent examples from the field. Additionally, we propose several enhancements to these existing models through a unified episodic memory model for ECA's. As part of our research into episodic memory models, we present a process for determining the prevalent contexts in the conversations obtained from the aforementioned interactions. The process presented demonstrates the use of statistical and machine learning services, as well as Natural Language Processing techniques to extract relevant snippets from conversations. Finally, mechanisms to store, retrieve, and recall episodes from previous conversations are discussed. A primary contribution of this research is in the context of contemporary memory models for conversational agents and cognitive architectures. To the best of our knowledge, this is the first attempt at providing a comparative summary of existing works. As implementations of ECAs become more complex and encompass more realistic conversation engines, we expect that episodic memory models will continue to evolve and further enhance the naturalness of conversations.

This dissertation is a multi-disciplinary study, which integrates a software engineering approach with instructional factors in the decision-making, analysis, design and development processes of instructional software. Software engineering models, tools and representations are used in the process of software construction. With reference to the fundamental characteristics of the software product, several disciplines and factors, from both instructional and computing perspectives are considered, and the most appropriate approach/es selected. Software engineering, instructional design and instructional theory are considered as pillars of courseware engineering. The object-oriented design paradigm and a prototyping life-cycle model are found to be most suitable for development of computer-aided instruction. The conceptual study is illustrated by prototype development of a component-based multi-activity practice environment in theoretical Computer Science. It offers perusal or practice, in various instructional modes, according to the user's preferred learning style or need. / Computing / M. Sc. (Information Systems)

En 1977, Freeman formalisait les premières mesures génériques d'Analyse de Réseaux Sociaux (ARS). Puis, les réseaux sociaux du Web " 2.0 " sont devenus planétaires (e.g., FaceBook, MSN). Cette thèse définit un modèle sémantique, non probabiliste et prédictif, pour l'analyse décisionnelle de réseaux sociaux professionnels et institutionnels. Ce modèle, en parallèle à la sociophysique de Galam, intègre des méthodes de traitement sémantique du langage naturel et d'ingénierie des connaissances, des mesures de sociologie statistique et des lois électrodynamiques, appliquées à l'optimisation de la performance économique et du climat social. Il a été développé et expérimenté dans le cadre du projet Socioprise, financé par le Secrétariat d'Etat à la prospective et au développement de l'économie numérique.

Un des objectifs fondamentaux du dialogue homme-machine est de se rapprocher du dialogue naturel en langage naturel, c'est-à-dire de permettre une interaction entre la machine et son utilisateur humain dans la langue de celui-ci (langage naturel), avec une structure d'échanges similaire à un dialogue humain (dialogue naturel). Les recherches impliquées se nourrissent de travaux linguistiques qui analysent la langue et de travaux pragmatiques qui analysent l'usage du langage en contexte. Deux facettes importantes de la pragmatique linguistique portent ainsi sur les phénomènes de référence, par exemple les désignations des objets accessibles dans le contexte situationnel, et sur les actes de langage, ou actes de dialogue, c'est-à-dire les actions communicatives effectuées par les énoncés constituant les tours de parole. Nous présentons nos travaux de modélisation et de formalisation de ces deux facettes, avec leur application au dialogue avec support visuel et au dialogue associant parole et gestes co-verbaux (dialogue multimodal). Un autre objectif du dialogue homme-machine est de mettre en oeuvre des méthodologies et des moyens, par exemple des architectures logicielles réutilisables, pour faciliter le développement de systèmes. Nous présentons nos réflexions et nos réalisations dans ce sens, à travers notamment notre participation à un ensemble de projets européens. Nous proposons enfin des perspectives de recherche qui visent à mieux intégrer au dialogue homme-machine des phénomènes linguistiques et pragmatiques telles que la saillance et l'ambiguïté.

Cette thèse se situe dans le domaine du Traitement Automatique des Langues et vise à optimiser la classification des documents dans les moteurs de recherche. Les travaux se concentrent sur le développement d'un outil de détection automatique des thèmes des documents (ATDS-fr). Utilisant peu de connaissances, la méthode hybride adoptée allie des techniques statistiques de segmentation thématique à des méthodes linguistiques identifiant des marqueurs de cohésion. Parmi eux, les chaînes de référence - séquence d'expressions référentielles se rapportant à la même entité du discours (e.g. Paul...il...cet homme) - ont fait l'objet d'une attention particulière, car elles constituent un indice textuel important dans la détection des thèmes (i.e. ce sont des marqueurs d'introduction, de maintien et de changement thématique). Ainsi, à partir d'une étude des chaînes de référence menée dans un corpus issu de genres textuels variés (analyses politiques, rapports publics, lois européennes, éditoriaux, roman), nous avons développé un module d'identification automatique des chaînes de référence RefGen qui a été évalué suivant les métriques actuelles de la coréférence.