No description available.

Cette thèse s’intéresse aux systèmes tutoriels intelligents (STI), un type d’environnement informatique pour l’apprentissage humain (EIAH) qui se distingue des autres (p. ex. les exerciseurs et les hypermédias éducatifs) en offrant un mécanisme d’évaluation plus sophistiqué. Parmi les différentes familles de STI, ce sont les STI « par traçage de modèle » (MTT) qui ont le plus fait leurs preuves. Les MTT sont critiqués, premièrement parce qu’ils évaluent l’apprenant de façon serrée (c.-à-d. qui positionne l’action de l’apprenant par rapport à une ou plusieurs méthodes pour effectuer la tâche), ce qui n’est possible que pour des tâches bien définies. Par conséquent, on leur reproche d’encourager un apprentissage superficiel. Deuxièmement, parce que les efforts de création qu’ils requièrent sont jugés prohibitifs, ce qui a mené à l’apparition d’autres familles de STI, comme les STI « par contraintes » et les STI « par traçage d’exemples » et ceux basés sur l’apprentissage automatique. Par cette thèse, nous voulons contribuer à renouveler l’intérêt pour les MTT en améliorant le rapport entre les efforts de création et l’efficacité potentielle des interventions, et en établissant plus clairement leur rôle pédagogique. Pour ce faire, nous proposons la plateforme Astus qui permet d’explorer l’espace qui existe entre les MTT créés avec les plateformes existantes, et des MTT dédiés ayant recours à des connaissances didactiques sophistiquées (p. ex. des dialogues) qui exigent des efforts de création encore plus importants. La plateforme Astus se distingue des plateformes existantes parce qu’elle génère des interventions plutôt que de recourir à des interventions prémâchées et qu’elle supporte les tâches s’effectuant dans des environnements qui ont une dimension physique. La génération des interventions dépend : d’un modèle de la tâche qui s’inscrit dans le paradigme du tuteur, c’est-à-dire qui représente une abstraction et une généralisation des instructions d’un tuteur humain; d’un modèle de l’UI qui permet des interventions riches comme une démonstration (c.-à-d. déplacements du pointeur et simulation des clics et des saisies); de langages dédiés et d’outils qui réduisent les efforts de création des auteurs; de mécanismes d’extension qui permettent d’adapter la génération en fonction d’une stratégie pédagogique particulière. Le paradigme du tuteur, parce qu’il favorise une communication transparente entre le système et l’apprenant, met en évidence les avantages et les désavantages de l’approche pédagogique des MTT, essentiellement une évaluation précise (c.-à-d. qui permet de produire des indices sur la prochaine étape et des rétroactions sur les erreurs), mais serrée. En s’inscrivant explicitement le paradigme du tuteur, entre autres en évitant de tirer profit de la nature de domaines particuliers ou de propriétés de tâches particulières pour assouplir l’évaluation, la plateforme Astus se démarque plus nettement des autres familles de STI que les autres MTT. Par conséquent, elle établit plus clairement le rôle pédagogique des MTT. Cinq expérimentations (menées par Luc Paquette) à petite échelle ont été réalisées auprès d’étudiants au baccalauréat au département d’informatique (un laboratoire pour la manipulation d’arbres binaires de recherche et un pour la conversion de nombres en virgule flottante). Ces expérimentations indiquent que les interventions générées sont efficaces. Au-delà de ces résultats, c’est le processus entourant ces expérimentations, parce qu’il est comparable au processus des chercheurs potentiellement intéressés par la plateforme Astus, qui montre que la version présentée dans cette thèse est plus qu’un prototype et qu’elle peut être utilisée à l’interne dans un contexte réel.

Recent trends in robotics have seen more general purpose robots being deployed in unstructured environments for prolonged periods of time. Such robots are expected to adapt to different environmental conditions, and ultimately take on a broader range of responsibilities, the specifications of which may change online after the robot has been deployed. We propose that in order for a robot to be generally capable in an online sense when it encounters a range of unknown tasks, it must have the ability to continually learn from a lifetime of experience. Key to this is the ability to generalise from experiences and form representations which facilitate faster learning of new tasks, as well as the transfer of knowledge between different situations. However, experience cannot be managed na¨ıvely: one does not want constantly expanding tables of data, but instead continually refined abstractions of the data – much like humans seem to abstract and organise knowledge. If this agent is active in the same, or similar, classes of environments for a prolonged period of time, it is provided with the opportunity to build abstract representations in order to simplify the learning of future tasks. The domain is a common structure underlying large families of tasks, and exploiting this affords the agent the potential to not only minimise relearning from scratch, but over time to build better models of the environment. We propose to learn such regularities from the environment, and extract the commonalities between tasks. This thesis aims to address the major question: what are the domain invariances which should be learnt by a long lived agent which encounters a range of different tasks? This question can be decomposed into three dimensions for learning invariances, based on perception, action and interaction. We present novel algorithms for dealing with each of these three factors. Firstly, how does the agent learn to represent the structure of the world? We focus here on learning inter-object relationships from depth information as a concise representation of the structure of the domain. To this end we introduce contact point networks as a topological abstraction of a scene, and present an algorithm based on support vector machine decision boundaries for extracting these from three dimensional point clouds obtained from the agent’s experience of a domain. By reducing the specific geometry of an environment into general skeletons based on contact between different objects, we can autonomously learn predicates describing spatial relationships. Secondly, how does the agent learn to acquire general domain knowledge? While the agent attempts new tasks, it requires a mechanism to control exploration, particularly when it has many courses of action available to it. To this end we draw on the fact that many local behaviours are common to different tasks. Identifying these amounts to learning “common sense” behavioural invariances across multiple tasks. This principle leads to our concept of action priors, which are defined as Dirichlet distributions over the action set of the agent. These are learnt from previous behaviours, and expressed as the prior probability of selecting each action in a state, and are used to guide the learning of novel tasks as an exploration policy within a reinforcement learning framework. Finally, how can the agent react online with sparse information? There are times when an agent is required to respond fast to some interactive setting, when it may have encountered similar tasks previously. To address this problem, we introduce the notion of types, being a latent class variable describing related problem instances. The agent is required to learn, identify and respond to these different types in online interactive scenarios. We then introduce Bayesian policy reuse as an algorithm that involves maintaining beliefs over the current task instance, updating these from sparse signals, and selecting and instantiating an optimal response from a behaviour library. This thesis therefore makes the following contributions. We provide the first algorithm for autonomously learning spatial relationships between objects from point cloud data. We then provide an algorithm for extracting action priors from a set of policies, and show that considerable gains in speed can be achieved in learning subsequent tasks over learning from scratch, particularly in reducing the initial losses associated with unguided exploration. Additionally, we demonstrate how these action priors allow for safe exploration, feature selection, and a method for analysing and advising other agents’ movement through a domain. Finally, we introduce Bayesian policy reuse which allows an agent to quickly draw on a library of policies and instantiate the correct one, enabling rapid online responses to adversarial conditions.

This study explores what motivates engineers in their early-mid career to change careers. It first establishes the definition of a career change from the perspective of those who have changed careers, and then examines what drives, influences and facilitates a career change, as well as how a career change is enacted physically and emotionally. This has been looked at through the contemporary lens of the ‘intelligent career framework’. This research adopted a qualitative, abductive approach following an initial inductive small-scale exploratory study. The fieldwork consisted of a pilot and main study using semi-structured interviews. For the main study, 22 interviews were conducted within one organisation in order to elicit the subjective experiences of engineers who had undertaken a career change. The findings show how the driving factors relate predominantly to knowing-why and knowing-where. The influencing and facilitating factors vary by individual, and relate to knowing-what, knowing-how knowing-when and knowing-whom. Six clusters of interacting factors were observed with knowing-why, knowing-how and knowing-when at the core. Whilst the process of career change was complex and long, differing pathways through which individuals changed careers were evident, as well as emotions that needed to be managed. This study contributes to knowledge in the area of contemporary career theory by exploring career change through a new lens: the intelligent career framework. It demonstrates how individuals use their ‘career capital’ to effect a career change and the ways in which the six knowings interact to bring about a career change. It extends the understanding of the process of career change and discovers some of the organisational factors that influence or facilitate individuals making a career change. All of these contributions address identifiable gaps in the literature.

International Telemetering Conference Proceedings / October 22-25, 1984 / Riviera Hotel, Las Vegas, Nevada / The processing power afforded by embedded microcomputers in state-of-the-art control applications offers the design engineer greatly expanded opportunities for improved ergonomic design, even without the use of “soft” actuator and/or display devices (which are sometimes undesirable and/or unacceptable). Especially important is the exploitation of software to simplify the hardware design while simultaneously implementing decision/mode logic that would be prohibitively expensive if done in hardware alone. The designer is often confronted with an ocean of possibilisties, and must make intelligent decisions in order to satisfy increasingly demanding applications and sophisticated users. The design decisions and resulting features and behaviors of an automatic antenna control unit are discussed from the operators point of view (black box), but also with the intention to detail some of the logic necessary to implement these features. This is prefaced by a discussion of the characteristics of the primary operator interface, the front panel, and the factors that influenced its design. Areas for future improvement of the design are also mentioned.

Both semantic web applications and individuals are in need of knowledge infrastructures that can be used in dynamic and distributed environments where autonomous entities create knowledge and build their own view of a domain. The prevailing view today is that the process of ontology evolution is difficult to monitor and control, so few efforts have been made to support such a controlled process formally involving several ontologies. The new paradigm we propose is to use an infrastructure mechanism that processes ontology change proposals from autonomous entities while maintaining user-defined consistency between the ontologies of these entities. This makes so called semantic autonomy possible. A core invention of our approach is to formalise consistency constraints as so called spheres of consistency that define 1) knowledge regions within which consistency is maintained and 2) a variable degree of proof-bounded consistency within these regions. Our infrastructure formalism defines a protocol and its computational semantics, as well as a model theory and proof theory for the reasoning layer of the mechanism. The conclusion of this thesis is that this new paradigm is possible and beneficial, assuming that the knowledge representation is kept simple, the ontology evolution operations are kept simple and one proposal is processed at a time.

published_or_final_version / Industrial and Manufacturing Systems Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / Industrial and Manufacturing Systems Engineering / Doctoral / Doctor of Philosophy

No description available.

No description available.