Empirical research provides evidence of strong interaction between cognitive and affective processes in the human mind. Education research proposes a model of constructive learning that relates cognitive and affective processes in an evolving cycle of affective states. Intelligent Tutoring Systems (ITSs) are capable of providing comprehensive cognitive support. Affective support in ITSs, however, is lagging behind; the in-depth exploration of cognitive and affective processes in ITSs is yet to be seen. Our research focuses on the integration of affective support in an ITS enhanced with an affective pedagogical agent. In our work we adopt the dimensional (versus categorical) view of emotions for modelling affective states of the agent and the ITSs users. In two stages we develop and evaluate an affective pedagogical agent. The affective response of the first agent version is based on the appraisal of the interaction state; this agent's affective response is displayed as affective facial expressions. The pilot study at the end of the first stage of the project confirms the viability of our approach which combines the dimensional view of emotions with the appraisal of interaction state. In the second stage of the project we develop a facial feature tracking application for real-time emotion recognition in a video-stream. Affective awareness of the second version of the agent is based on the output from the facial feature tracking application and the appraisal of the interaction state. This agent's response takes the form of affectoriented messages designed to interrupt the state of negative flow. The evaluation of the affect-aware agent against an unemotional affect-unaware agent provides positive results, thus confirming the superiority of the affect-aware agent. Although the uptake of the agent was not unanimous, the agent established and maintained good rapport with the users in a role of a caring tutor. The results of the pilot study and the final evaluation validate our choices in the design of affective interaction. In both experiments, the participants appreciated the addition of audible feedback messages, describing it as an enhancement which helped them save time and maintain their focus. Finally, we offer directions for future research on affective support which can be conducted within the framework developed in the course of this project.

Intelligent tutoring systems and computer-supported collaborative environments have been designed to enhance human learning in various domains. While a number of solid techniques have been developed in the Artificial Intelligence in Education (AIED) field to foster human learning in fundamental science domains, there is still a lack of evidence about how to support learning in so-called ill-defined domains that are characterized by the absence of formal domain theories, uncertainty about best solution strategies and teaching practices, and learners' answers represented through text and argumentation. This dissertation investigates how to support students' learning in the ill-defined domain of professional ethics through a computer-based learning system. More specifically, it examines how to support students in the analysis of case studies, which is a common pedagogical practice in the ethics domain. This dissertation describes our design considerations and a resulting system called Umka. In Umka learners analyze case studies individually and collaboratively that pose some ethical or professional dilemmas. Umka provides various types of support to learners in the analysis task. In the individual analysis it provides various kinds of feedback to arguments of learners based on predefined system knowledge. In the collaborative analysis Umka fosters learners' interactions and self-reflection through system suggestions and a specifically designed visualization. The system suggestions offer learners the chance to consider certain helpful arguments of their peers, or to interact with certain helpful peers. The visualization highlights similarities and differences between the learners' positions, and illustrates the learners' level of acceptance of each other's positions. This dissertation reports on a series of experiments in which we evaluated the effectiveness of Umka's support features, and suggests several research contributions. Through this work, it is shown that despite the ill-definedness of the ethics domain, and the consequent complications of text processing and domain modelling, it is possible to build effective tutoring systems for supporting students' learning in this domain. Moreover, the techniques developed through this research for the ethics domain can be readily expanded to other ill-defined domains, where argument, qualitative analysis, metacognition and interaction over case studies are key pedagogical practices.

Emotions play a significant role in many human mental activities, including decision-making, motivation, and cognition. Various intelligent and expert systems can be empowered with emotionally intelligent capabilities, especially systems that interact with humans and mimic human behaviour. However, most current methods in affect recognition studies use intrusive, lab-based, and expensive tools which are unsuitable for real-world situations. Inspired by studies on keystrokes dynamics, this thesis investigates the effectiveness of diagnosing users’ affect through their typing behaviour in an educational context. To collect users’ typing patterns, a field study was conducted in which subjects used a dialogue-based tutoring system built by the researcher. Eighteen dialogue features associated with subjective and objective ratings for users’ emotions were collected. Several classification techniques were assessed in diagnosing users’ affect, including discrimination analysis, Bayesian analysis, decision trees, and neural networks. An artificial neural network approach was ultimately chosen as it yielded the highest accuracy compared with the other methods. To lower the error rate, a hierarchical classification was implemented to first classify user emotions based on their valence (positive or negative) and then perform a finer classification step to determining which emotions the user experienced (delighted, neutral, confused, bored, and frustrated). The hierarchical classifier was successfully able to diagnose users' emotional valence, while it was moderately able to classify users’ emotional states. The overall accuracy obtained from the hierarchical classifier significantly outperformed previous dialogue-based approaches and in line with some affective computing methods.

Thesis (Ph. D.)--University of Canterbury, 2007. / Typescript (photocopy). Includes bibliographical references. Also available via the World Wide Web.

Thesis (M.S.)--State University of New York at Binghamton, Department of Computer Science, Thomas J. Watson School of Engineering and Applied Science, 2007. / Includes bibliographical references.

Thesis (M. Sc.)--University of Canterbury, 2008. / Typescript (photocopy). "January 2008." Includes bibliographical references (p. [121]-128). Also available via the World Wide Web.

Thesis (M.S.)--Worcester Polytechnic Institute. / Keywords: Intelligent Tutoring Systems; Model Tracing; Machine Learning; Artificial Intelligence; Programming by Demonstration. Includes bibliographical references.

Thesis (M. Sc.)--University of Canterbury, 2009. / Typescript (photocopy). "September 2009." Includes bibliographical references (leaves 89-96). Also available via the World Wide Web.

Thesis (Ph. D.)--University of Canterbury, 2006. / Typescript (photocopy). Includes bibliographical references (p. 246-255). Also available via the World Wide Web.

Οι εξατομικευμένες ασκήσεις είναι ένα χαρακτηριστικό γνώρισμα στην προώθηση της σύγχρονης εκμάθησης στον ιστό. Ένα σημαντικό πρόβλημα που συναντάται στα συστήματα μάθησης στον ιστό είναι ο μικρός αριθμός διαφορετικών ερωτήσεων-ασκήσεων. Αυτό οδηγεί σε μια ανακύκλωση των ίδιων ερωτήσεων-ασκήσεων για κάθε σπουδαστή και όσον αφορά την εξάσκησή του και όσον αφορά στην αξιολόγηση του. Έτσι, ο αριθμός των διαθέσιμων ερωτήσεων-ασκήσεων είναι χαρακτηριστικά ανεπαρκής για την αξιολόγηση του επιπέδου της γνώσης και την καθοδήγηση σε περαιτέρω μελέτη. Στην μεταπτυχιακή αυτή εργασία παρουσιάζεται το σύστημα SchoolMath, το οποίο μπορεί να παράγει πρακτικά ένα πολύ μεγάλο αριθμό παραμετροποιημένων ασκήσεων στα μαθηματικά, πιο συγκεκριμένα ασκήσεων που αναφέρονται στην επίλυση συστημάτων δύο εξισώσεων με δύο αγνώστους με τη βοήθεια των οριζουσών. Ο χρήστης-μαθητής (ή και καθηγητής) μπορεί να καθορίσει και το επίπεδο δυσκολίας της παραγόμενης άσκησης (εύκολη, μέτρια δύσκολη). Το σύστημα εκτός από τη παραγωγή ασκήσεων μπορεί να αποτελέσει ένα σημαντικό εργαλείο-βοήθημα για την εξατομικευμένη μάθηση των μαθητών, για την εκμάθηση επίλυσης συστημάτων με τη βοήθεια των οριζουσών στον υπολογιστή του. Σε κάθε βήμα του ο μαθητής μπορεί να έχει ανατροφοδότηση δύο επιπέδων, εφ΄ όσον τη ζητήσει. Θα μπορούσε επίσης να είναι ένα βοήθημα για την δημιουργία ασκήσεων για τον συνάδελφο εκπαιδευτικό των μαθηματικών της δευτεροβάθμιας εκπαίδευσης. Το αντικείμενο αυτό των μαθηματικών είναι από τα πρώτα που παρουσιάζονται στο βιβλίο της άλγεβρας για μαθητές της Α΄ Λυκείου, αφού έχουν διδαχθεί στην Γ΄ Γυμνασίου την επίλυση του συστήματος των δύο εξισώσεων με δύο αγνώστους με τη μέθοδο της αντικατάστασης και τη μέθοδο των αντιθέτων συντελεστών. Υποτίθεται βέβαια ότι ο μαθητής έχει διδαχθεί στην τάξη την έννοια της ορίζουσας και τον τρόπο λύσης (με ορίζουσες) ενός 2x2 συστήματος. Έτσι, έχουμε σχεδιάσει και υλοποιήσει ένα σύστημα εκμάθησης για τα μαθηματικά σε ένα τομέα όπου δεν έχει ξανά ερευνηθεί ή τουλάχιστον ακόμη δεν έχει παρουσιαστεί κάτι ανάλογο για να μπορέσουμε να το συγκρίνουμε άμεσα. Τα περισσότερα συστήματα που έχει γίνει έρευνα ασχολούνται με θέματα του τομέα της πληροφορικής, που έχει αρκετές διαφορές στο τρόπο λύσης των ασκήσεων σε σχέση με αυτό των μαθηματικών. / Individualised exercises are a characteristic trait in the promotion of modern learning on the web. An important problem met in the web-based tutoring systems is the small number of different questions-exercises that they provide. This leads to a recycling of the same questions-exercises for each student with regards to both their exercising and evaluation. Thus, the number of available questions-exercises is remarkably inadequate for evaluating their knowledge level and for guiding them in further study. In this Master Thesis, system “SchoolMath” is presented, which can generate a practically very large number of parameterised exercises in mathematics, more specifically exercises related to resolution of a system consisting of two equations with two unknown variables with the method of determinants. The user-student (or user-tutor) can specify the level of difficulty of the produced exercise (easy, medium, difficult). The system, apart from exercises generation, may constitute a tool for helping students personalised learning, by exercising with resolution of two-equations systems with the determinants method on their personal computer. In each step of the process, the student can have feedback from the system of two levels, if he/she asks for it. Also, it could be an aid for creation of exercises for the teachers of mathematics of secondary education. This topic of mathematics is from the first ones presented in the book of algebra for first year students of Lyceum, after they have been taught the resolution of a system of two equations with two unknown variables with the methods of replacements and opposite factors, in the third year of High School. It is supposed, of course, that the student has been taught the notion of determinant and how to solve a 2x2 equation system in the classroom. Thus, we have designed and implemented a system for learning in mathematics, related to a topic that has not been investigated or at least there is not something similar in the existing literature, as far as we are aware of. Most of existing systems are concerned with topics related to informatics, which have a different way of approaching the solution of problems compared to that in mathematics.