Search

The paper presents an approach for the dynamic generation of a complex structure of personalized web pages for learning purposes, reflecting the ontology of the considered domain. The need of assuring a holistic character for the body of knowledge induced in the learner's mind is emphasized. This is very important in the learning processes, especially nowadays, in the context of the huge amount of information available on the web and of its permanent evolution. The approach permits the adaptation of the content of the generated web pages to the incoming information from the web. New information is extracted, annotated and coherently integrated in the body of knowledge in order to keep the holistic character of the body of knowledge. Personalization is achieved by filtering the semantic network according to the learner model, which keeps the list of concepts known or unknown by the learner. The approach was used in an EU INCO Copernicus project for computer aided language learning. [ABSTRACT FROM AUTHOR]

The integration of GRID and MAS (Multi-Agents Systems) is an active research topic. We have recently proposed the Agent-Grid Integration Language, to describe a service-based integration of GRID and MAS models. However, the complexity of the mutual integration aspects leads us to define a rigorous way to formalize the key concepts, their relations and the integration rules by means of an ontology. With this ontology, we can describe the elements and their composition that occur in various service exchange scenarios with agent on the Grid. The ontology could be used both to model the behaviour of GRID-MAS integrated systems and to check the consistency of these systems and their instances. A concrete scenario is illustrated. [ABSTRACT FROM AUTHOR]

Intelligent tutoring Systems (ITSs) adapt content and activities with the goals of being both effective and efficient instructional environments. They have goals for students to be challenged and guided in an optimal way-without being too overwhelmed with difficult material or too bored with easy or repetitive material. We propose a particular definition of the zone of proximal development (ZPD) as a general way to describe what all ITSs try to do, and we propose a foundational analysis of instructional adaptivity, student modeling, and system evaluation in terms of the ZPD. We give an operational definition of the ZPD and give an example of its use, and summarize how instructional methods such as scaffolding can be used to maintain ZPD-learning. We also explain how our definition of the ZPD can lead to a more complete model for efficient and effective instruction than common mastery learning criteria. [ABSTRACT FROM AUTHOR]

In this paper, we describe an Intelligent Learning Environment which realizes the learning by problem posing. In the learning by problem posing, a learner poses problems through the interface provided by the ILE. The ILE has a function to diagnose the problems posed by the learner. By using the results of the diagnosis, the ILE helps the learner to correct wrong problems, or leads her/him in the next step of problem posing. We have developed the ILE. In this paper, we also report an experimental evaluation in an elementary school. By using the result of the experiment, we examined (1) usability of the ILE and (2) effectiveness of the ILE. In the ILE, the interface was implemented in Java, and the diagnosis module was implemented in Prolog. The current environment deals with simple arithmetical word problems solved by an addition or a subtraction. [ABSTRACT FROM AUTHOR]

The effectiveness of a Navigation Agent has been evaluated, which determines a path through course material tailored to the individual student. Lesson selection by the agent showed an advantage for students who were unable to monitor their own learning process. Initiative by the agent -where the agent also chooses the moment for selecting another lesson- reduced the time needed to finish the task by stimulating the students to study the lessons in a balanced way. The experiment also provides evidence that a mixed locus of control is preferable to either the student or the agent exclusively taking the initiative and selecting lessons. [ABSTRACT FROM AUTHOR]

We investigated college students' perceptions of a diverse sample of animated conversational agents. We also examined the pedagogical efficacy of those agents. We found that people perceive differences among the agents on several dimensions, such as likeability, and that the agents differ in pedagogical efficacy. However, none of the characteristics that we measured accounted for differences in pedagogical efficacy across the agents. We discuss implications for the field of agent studies with particular emphasis on the creation of peda-gogically effective conversational agents and suggest directions for future research. [ABSTRACT FROM AUTHOR]

The Human Use Regulatory Affairs Advisor (HURAA) is a web-based help facility that provides training on the ethical use of human subjects in research, and that serves as an information retrieval system on ethical policies. The content for HURAA is derived from United States Federal agency documents and regulations. HURAA has a number of components that go beyond conventional page-turning or hypertext systems, including (1) an animated conversational agent that serves as a navigational guide for the web facility, (2) an enhanced multimedia introduction, (3) lessons with case-based and explanation-based reasoning, and (4) document retrieval through natural language queries or a Point & Query facility. The effectiveness of HURAA was tested on a small sample of participants (N = 18) who were assigned to either the full HURAA version or a conventional computer-based training version. [ABSTRACT FROM AUTHOR]

There is a growing body of evidence that supports the claim that affect plays a critical role in decision-making and performance as it influences cognitive processes [1], [2], [3]. Despite this body of research the role and function of affect is not generally recognized by the disciplines that address the broad issues of understanding complex systems and complex behavior, especially in the presence of learning. The innovative models and theories that have been proposed to facilitate advancement in the field of human-computer interaction (HCI) tend to focus exclusively on cognitive factors. Consequently, the resulting systems are often unable to adapt to real-world situations in which affective factors play a significant role. We propose several new models for framing a dialogue leading to new insights and innovations that incorporate theories of affect into the design of (affect-sensitive) cognitive machines. [ABSTRACT FROM AUTHOR]

We present a probabilistic model that assesses student emotional reaction during interaction with an educational game. Following a well-known cognitive theory of emotions (the OCC theory), the model predicts a student's emotional state by assessing the student's appraisal of her interaction with the game, in light of the student's goals and personality. We illustrate how the model relies on a Dynamic Decision Network that is based on both the OCC theory and observations from two user studies. [ABSTRACT FROM AUTHOR]

The ability to detect the students' motivational state during an instructional interaction can bring many benefits to the performance of an Intelligent Tutoring System (ITS). In this paper we present an empirical study which provided us with a considerable amount of knowledge regarding motivation diagnosis. We show how this knowledge was formalised in order to create a set of motivation diagnosis rules that can be incorporated into a prototype tutoring system. We also briefly present how these motivation diagnosis rules were evaluated in another study. [ABSTRACT FROM AUTHOR]

In this paper we discuss how collaboration issues are addressed in the context of Teatrix, a virtual environment aiming at providing the children with the means for collaboratively creating a story on a virtual stage. The children create stories using a set of pre-defined scenes and dramatis personae — characters that have specific roles in the play, which they control to a certain extent). Each child expects the story to evolve in reaction to her/his character's actions and the overall story emerges from the collaboration of different children. Based on this premise we conducted some evaluation experiences, in order to assess to what extent Teatrix promoted collaboration and what types of collaboration were particularly elicited by the application. [ABSTRACT FROM AUTHOR]

This paper describes StoryStation, an intelligent tutoring system designed to give ten to twelve year old children feedback on their creative writing. The feedback is presented via eight animated interface agents. Each agent gives a different sort of support to the writer including: a thesaurus, a dictionary, feedback on vocabulary and characterisation, help with spelling, help with plot structure, example stories to read and help with the interface itself. This paper focuses on the strategies for generating feedback to the children and discusses some issues in presenting this feedback through the interface agents. [ABSTRACT FROM AUTHOR]

This paper describes a pedagogical agent that helps children to learn to author structured presentations about explanations of concepts. Using a Rhetorical Structure Theory analysis of a source Web page, the agent performs pedagogical tasks to support the user's understanding of rhetorical relations, stimulates reflection about the relations between the structure of the original text and the structure of the presentations, and suggests ways to improve the user's performance. Upon completion of the authoring, the presentations are organized into coherent structures that can be performed by animated characters, or Digital Puppets, in a learning-by-teaching classroom context. [ABSTRACT FROM AUTHOR]

The REDEEM authoring tool allows teachers to create simple ITSs from existing CBT by imposing their pedagogical preferences about how students should best be taught. In this study we compared learning outcomes of 86 14-16 yr old students who studied genetics with CBT and with ITSs constructed for them by their class teacher who had expert knowledge of both the topic and these students. We found that performance of the students improved from pre-test to post-test but learning outcomes were not influenced by type of learning environment. Detailed analysis of participants' performance suggested that REDEEM could enhance knowledge by directly questioning students but that features such as differentiation did not impact upon performance. We discuss possible interpretations for these results in the light of evaluation issues for ITS authoring tools. [ABSTRACT FROM AUTHOR]

StoryML is an XML-based representation of metadata elements connected to collaboratively written stories. The StoryML specification gives means to interpret and specify the core characteristics of stories. Hence, StoryML supports the functionality of information retrieval, filtering and adaptive representation of stories. These intelligence properties make collaboratively written stories a significant platform for truly activating, open learning environments. [ABSTRACT FROM AUTHOR]

One of the traditional goals of an Intelligent Tutoring System is to provide domain content for learners that is appropriate to their needs. A key component of this knowledge is the sequence in which instructional activities are performed and so unsurprisingly this is often a key task in ITS authoring. In order to understand this process better it is important to have accurate quantitative ways of classifying the difference between alternative sequences of ITS material. Edit distance algorithms provide a useful way of capturing this knowledge but suggest that weights need to be careful adjusted to capture important aspects of ITS sequences. The weighted algorithm is illustrated with examples from three studies using the REDEEM ITS authoring tool. This technique has allowed us to compare authors' sequences in a way that is robust, quantifiable and that provides insights into ITS authors' pedagogical principles. [ABSTRACT FROM AUTHOR]

Our goal is to find a methodology for directing development effort in an intelligent tutoring system (ITS). Given that ITS have several AI reasoning components, as well as content to present, evaluating them is a challenging task. Due to these difficulties, few evaluation studies to measure the impact of individual components have been performed. Our architecture evaluates the efficacy of each component of an ITS and considers the impact of a particular teaching goal when determining whether a particular component needs improving. For our AnimalWatch tutor, we found that for certain goals the tutor itself, rather than its reasoning components, needed improvement. We have found that it is necessary to know what the system's teaching goals are before deciding which component is the limiting factor on performance. [ABSTRACT FROM AUTHOR]

Since the development of the first digital computer in the 1940s, the notion of computer intelligence has received considerable attention from computer scientists, philosophers, and psychologists. The question of whether it is possible to create a computer program that possesses human intelligence has spurred much debate. Turing (1950) argued that computers are not capable of thinking and provided several theological, psychological, and sociological arguments in support of his position. To determine a computer program's intelligence, Turing proposed several benchmark methods. One such method requires humans to decide whether they are interacting with an actual computer program or another human via computer mediation. According to Turing, a computer could be described as intelligent if it could deceive a human into believing that it was human. The two studies presented here were designed to determine whether AutoTutor could pass a variation of the Turing test, the Bystander Turing Test. The subsequent sections of this paper address the following: (1) the AutoTutor system, (2) the Bystander Turing Test, (3) the two empirical studies, and (4) the conclusions of the studies. [ABSTRACT FROM AUTHOR]

Our research concerns distance learning (DL). We are interested with distributed collaborative learning. In this approach, it is important to have indicators permitting the appreciation of durability and the evolution of groups involved. We think that actors responsible for the organisation and the working of groups (tutor for each group and coordinator of the DL session for all groups and its progress in general) can from the types of interactions and their amounts, get revealing elements permitting them to appreciate the state of a group and its evolution. From the analysis of interactions seen during a distance learning experimentation that we led, we show here that the disappearance of a group as we observed could be discerned practically in real time. It justifies for us, the necessity to set up in distance learning environments, agents capable of assisting the coordinator of the training and the tutors in their tasks. [ABSTRACT FROM AUTHOR]

DEMIST is a multi-representational simulation environment that supports understanding of the representations and concepts of population dynamics. We report on a study with 18 subjects with little prior knowledge that explored if DEMIST could support their learning and asked what decisions learners would make about how to use the many representations that DEMIST provides. Analysis revealed that using DEMIST for one hour significantly improved learners' understanding of population dynamics though their knowledge of the relation between representations remained weak. It showed that learners used many of DEMIST's features. For example, they investigated the majority of the representational space, used dyna-linking to explore the relation between representations and had preferences for representations with different computational properties. It also revealed that decisions made by designers impacted upon what is intended to be a free discovery environment. [ABSTRACT FROM AUTHOR]

This paper intends to present the design of an intelligent assessment system, which attempts to assess the student in arithmetic word problem solving. During assessment, the system keeps track of the aptitudes, which the student shows, concerning the answers he/she gives to special types of problems and in the same time it observes aspects of the student's motivational state. More precisely, motivation aspects deal with a) the effort the student shows in solving the different types of problems and b) with the independency and the confidence that characterize his/her behaviour. The system tries to adapt itself according to the above information, in order to motivate the student, offering her the appropriate help and the possibility to follow an individualized way through the objective items of the assessment. [ABSTRACT FROM AUTHOR]

The benefits of using Model-Based Reasoning for domain modeling are several-fold. We analyze these benefits and illustrate them in the context of a Web-based Intelligent Tutoring System. The system is designed to teach students to analyze and debug C++ programs for semantic and run-time errors. We have evaluated one instance of the Model-Based tutor, which deals with debugging pointers in C++, in several sections of Computer Science II course. We will present the results of these evaluations, which confirm the learnability of Model-Based tutors. [ABSTRACT FROM AUTHOR]

We describe how Computer Algebra Systems (CASs) can be used as cognitive tools in a learning environment. In particular, we show how a CAS is employed in ActiveMath, how different types of CAS-exercises are designed, and how feedback can be produced with the help of the very same CAS. We report the results of a first preliminary formative evaluation of ActiveMath' CAS-exercises in a university course and some modifications of ActiveMath caused by this evaluation. [ABSTRACT FROM AUTHOR]

Designing an Interactive Learning Environment (ILE) requires work within a multi-disciplinary team [9]. In this paper, we describe the designing process of the ILE AMBRE, the purpose of which is to teach methods. We show how, based on didactic work, we chose to represent the expert knowledge to be taught to the learner through Artificial Intelligence techniques. We also describe how we adjusted the Case-Based Reasoning paradigm (CBR), derived from Artificial Intelligence, to teaching abstract knowledge based on problem classes. Last we describe a Cognitive Psychology experiment we conducted in order to test the appropriateness of the thus designed EIAH to the goals of the AMBRE project. [ABSTRACT FROM AUTHOR]

Software scaffolding has been successfully employed within educational technology to help bridge the recognition-production gap between what learners want to achieve and what they are able to effect themselves without assistance. Such work has however concentrated on scaffolding the learner at the domain level with less attention to the potential for providing explicit support at the Metacognitive level. Evidence from previous work has shown that less able and less knowledgeable learners are especially ineffective at selecting appropriately challenging tasks and seeking appropriate qualities and quantities of support and guidance [1, 2]. But how can we make learners more effective at reflecting on their own needs, at seeking appropriate challenges and appropriate support? We have used a participatory design approach to assist young learners in the design process so that they can, in turn, assist us as we develop Metacognitive scaffolding strategies. These strategies have been implemented in Ecolab II. Early results are encouraging and suggest that low ability children can too be scaffolded to greater success. [ABSTRACT FROM AUTHOR]

Current systems aiming at engaging students in Self-Regulated Learning processes are often prompt-based and domain-dependent. Such metacognitive prompts are either difficult to interpret for novices or ignored by experts. Although domain-dependence per se cannot be considered as a drawback, it is often due to a rigid structure which prevents from moving to another domain. We detail here Apex, a two-loop system which provides texts to be learned through summarization. In the first loop, called Reading, the student formulates a query and is provided with texts related to this query. Then the student judges whether each text presented could be summarized. In the second loop, called Writing, the student writes out a summary of the texts, then gets an assessment from the system. In order to automatically perform various comprehension-centered tasks (i.e., texts that match queries, assessment of summaries), our system uses LSA (Latent Semantic Analysis), a tool devised for the semantic comparison of texts. [ABSTRACT FROM AUTHOR]

This paper introduces, EDUCE, an ITS that utilises individual learning characteristics to generate presentations in diverse and sensitive ways. In EDUCE, the pedagogical framework classifies the educational content and the learner characteristics in the student model along two dimensions: Gardner's Multiple Intelligences and Bloom's learning goals. The two dimensions represent the philosophical underpinning in the design of instructional strategies and for understanding the student behaviour. It is through the provision of a variety of instructional strategies, that EDUCE aims to motivate and engage the learner. This paper describes the principles, architecture, design and implementation of EDUCE. It shows how educational theory may underpin the design of an ITS and how a pedagogical component that accommodates learning characteristics may be incorporated into an ITS. It also shows how to develop a mechanism by which the learner can choose between alternative instructional approaches. [ABSTRACT FROM AUTHOR]

SIETTE is an efficient web-based implementation of a Computer Adaptive Test. The inference machine used is based on Item Response Theory. New enhances in the evaluation mechanisms, question selection and finalization criteria have been introduced. New evaluation mechanism allows giving structured knowledge estimation about all topics evaluated in a test. Question selection criteria are able to automatically select a balanced number of items from all topics, so teachers do not need to accomplish this task manually. This paper shows that SIETTE can successfully be integrated into web-based Intelligent Tutoring Systems with structured curriculum, in order to make initial estimations of the student's knowledge level, or even to update the student's model after his exposition to instructional components. [ABSTRACT FROM AUTHOR]

The initialization of the student model in an intelligent tutoring system is a crucial issue. It is not realistic to assume that each new student has the same prior knowledge concerning the topic being taught, be it nothing or some "standard" prior knowledge. We introduce CLARISSE, which is a novel categorization method. We illustrate this tool with the identification of categories among students for QUANTI, an intelligent tutoring system for the teaching of quantum information processing. In order to classify a new learner, CLARISSE generates an adaptive pre-test that can identify with high accuracy the learner's category after very few questions. [ABSTRACT FROM AUTHOR]

This research explores the use of software to move higher education towards more active student learning including inquiry, problem-based and cooperative learning. Active students ask their own questions, engage in hypothesis generation, make and test predictions about theories, are involved in reasoning and address their own misunderstandings. These activities place heavy demands on the faculty and are rarely allowed in large classrooms. The challenge is to support this type of inquiry learning without also adding and excessive burden to faculty time. We have built and evaluated software that supports inquiry learning in classrooms. The software supports students to reason about a phenomenon, make mistakes and monitor their own scientific processes. Intelligent tutoring and a discovery approach guide students' inquiry in problem-cases. The software provides some of the efficiency needed to make inquiry-oriented instruction more widely available and enhance problem-solving activities during classtime. We are expanding this model for inquiry-based learning across three domains, several institutions and teaching style. Student performance will be measured, providing empirical evidence for the portability of inquiry-oriented instruction for college-level classrooms. We will evaluate the comparative results concerning the differential effects of these interventions in each environment. [ABSTRACT FROM AUTHOR]

In this paper, we present an approach to hierarchical knowledge representation for the student's evaluation in propositional logic. The hierarchical evaluation consists in assessing the student's state of knowledge at several levels of granularity. The relevance of the method is justified by the need for a precise and flexible diagnosis of the learner's skills in a given domain. For that purpose, we shall model the propagation of the evaluation from a specific level of knowledge content to more general levels, using Bayesian inferences and neural networks classifications. [ABSTRACT FROM AUTHOR]

Open learning environments provide a large amount of freedom and control, which can be beneficial for students who are able to explore the environment effectively, but can also be problematic for those who are not. To address this problem, we have designed a student model that allows an open learning environment to provide the students with tailored feedback on the effectiveness of their exploration. The model, which uses Bayesian Networks, was created by an iterative design and evaluation process. The successive evaluations were used to improve the model and to provide initial support for its accuracy and usefulness. [ABSTRACT FROM AUTHOR]

Design and simulation environments are tools offering contrasted benefits for discovery learning in multiple domains. However, when these domains are complex the problems that the students have to solve during their learning justify the necessity to carry out the design activity itself in groups. From the perspective of modelling and simulation in collaboration, we present DomoSim-TPC, a system supporting collaborative learning in Domotics1. The system provides shared workspaces integrating tools for domain problem solving with generic but customized tools to support collaborative decisions and discussions. The potential of this synthesis to support learning tasks in modelling is described. We outline the mechanisms offered by the system and discuss our solution compared to other relevant systems. [ABSTRACT FROM AUTHOR]

This paper describes our analysis of the literature on tutorial dialogues and presents a compilation of useful principles that students and teachers typically follow in making tutoring interactions successful. The compilation is done in the context of making use of those principles in building knowledge acquisition interfaces since acquisition interfaces can be seen as students acquiring knowledge from the user. We plan to use these ideas in our future work to develop more proactive and effective acquisition interfaces. [ABSTRACT FROM AUTHOR]

This paper's aim is to give a conceptual framework about the conception of a Computer Assisted Language Learning environment based on the communicative and actional approaches. In this environment, the learner is implicated in a task that he/she has to complete by interacting with a partner using his/her language competencies. Given a task, different ways exist to accomplish it and the different competencies of the learner can then be solicited with different types of interaction. Works accomplished in human computer dialogue give models to manage this interaction. [ABSTRACT FROM AUTHOR]

Studies have shown that when learners are working in groups, even when these groups are working remotely, they adopt a certain social behavior. It would seem particularly interesting to be able to determine these types of behavior automatically using a computer. Indeed, such information can be useful for the tutor working with the learners, for the learners themselves as well as for a computer system which aims to help these same learners. This paper presents an approach for automatic analysis of learners' social behavior during synchronous conversations. This approach is based on the specification and design of a textual conversation tool semi-structured by communicative acts and a computer agent to analyze the conversations automatically. [ABSTRACT FROM AUTHOR]

Recent studies reveal that human tutoring sessions do not always end when the student has solved a problem. Instead, tutors and students frequently use a post-practice discussion to bring new topics to the table or to continue problem-solving discussions. One of the main roles of post-practice dialogues is to support transfer—that is, the student's ability to apply concepts and adapt familiar solution strategies to unfamiliar problems. Several developers of intelligent tutoring systems have implemented post-practice modules, with similar aims. However, in contrast to the integrated instructional planning that human tutors apparently perform, automated planning of reflective activities is typically done independently of instructional planning during problem solving. We present a framework for describing reflective plans that are distributed between problem solving and debrief and evidence that reflective discussions support transfer in elementary mechanics. [ABSTRACT FROM AUTHOR]

The emphasis on building co-operative/collaborative environments has brought out the matter of group interactions, together with the issue of conflicts. If well employed, conflicts can trigger cognitive changesa. In this paper, we present the results obtained in the experimental studies carried out to evaluate our framework for detecting and mediating Meta- Cognitive conflicts. The theoretical framework developed enables a computational system to analyse the ongoing interaction, to detect and to mediate conflicts. MArCo, our prototype, shows how our conceptual framework has been put to use. The results obtained in the experimental studies lead us to think that the approach is an interesting one, and should be developed further. [ABSTRACT FROM AUTHOR]

Symbolization is the ability to translate a real world situation into the language of algebra. We believe that symbolization is the single most important skill students learn in high school algebra. We present research on what makes this skill difficult and report the discovery of a "hidden" skill in symbolization. Contrary to past research that has emphasized that symbolization is difficult due to both comprehension difficulties and the abstract nature of variables, we found that symbolization is difficult because it is the articulation in the "foreign" language of "algebra". We also present Ms. Lindquist, an Intelligent Tutoring System (ITS) designed to carry on a tutorial dialog about symbolization. Ms. Lindquist has a separate tutorial model encoding pedagogical content knowledge in the form of different tutorial strategies, which were partially developed by observing an experienced human tutor. We discuss aspects of this human tutor's method that can be modeled well by Ms. Lindquist. Finally, we present an early formative showing that students can learn from the dialogs Ms. Lindquist is able to engage student in. Ms. Lindquist has tutored over 600 students at www.AlgebraTutor.org. [ABSTRACT FROM AUTHOR]

This paper presents an instructional assistant agent for FLE2 - a distributed collaborative learning environment. We discuss the role of the instructional assistant agent and how it supports both the instructor and students in a distributed collaborative knowledge building process. We emphasize the supplementary role of the instructional assistant agent, which, on one hand, observes the distributed collaborative learning process and computes statistics for viewing, and on the other hand, detects possible problems and presents them to the instructor so that the instructor, if desired, gives feedback to the students so that they themselves can regulate the collaboration. [ABSTRACT FROM AUTHOR]

We compared two automated approaches to teaching distinguishing, a fundamental skill of case-based reasoning that involves assessing the relevant differences among cases in a context-sensitive way. The approaches are implemented in two versions of CATO, an ITS designed to teach law students basic skills of case-based legal argument. The original version of CATO employed a didactic explanatory dialogue. The newer version, CATO-Dial, teaches the same skill with a simulated dialectic argument in a courtroom setting. Our hypothesis was that students would learn better by engaging in the simulated argument than by receiving interactive explanation. We showed that students in the dialectic argument simulation group performed significantly better on certain sections of the post-test aimed at assessing transfer of their skills of distinguishing. [ABSTRACT FROM AUTHOR]

In this paper we explore the problem of selecting appropriate Knowledge Construction Dialogues (KCDs) for the purpose of encouraging students to include important points in their qualitative physics explanations that are missing. We describe a hybrid symbolic/statistical approach developed in the context of the Why2 conceptual physics tutor (Vanlehn et al., 2002). Our preliminary results demonstrate that our hybrid approach outperforms both the symbolic approach and the statistical approach by themselves. [ABSTRACT FROM AUTHOR]

Managing tutorial dialogue is an intrinsically complex task that is only partially covered by current models of dialogue processing. After an analysis of such models identifying their strengths and weaknesses, we propose a flexible, modular, and thus re-usable computational framework, centered around a 3-tier dialogue planning architecture. [ABSTRACT FROM AUTHOR]

Research Methods Tutor is a new intelligent tutoring system created by porting the existing implementation of the AutoTutor system to a new domain, Research Methods in Behavioural Sciences, which allows more interactive dialogues. The procedure of porting allowed for an evaluation of the domain independence of the AutoTutor framework and for the identification of domain related requirements. Specific recommendations for the development of other dialogue-based tutors were derived from our experience. [ABSTRACT FROM AUTHOR]

Research on intelligent tutoring systems has not leveraged general models of collaborative discourse, even though tutoring is inherently collaborative. Similarly, research on collaborative discourse theory has rarely addressed tutorial issues, even though teaching and learning are important components of collaboration. We help bridge the gap between these two related research threads by presenting a tutorial agent, called Paco, that we built using a domain-independent collaboration manager, called Collagen. Our primary contribution is to show how a variety of tutorial behaviors can be expressed as rules for generating candidate discourse acts in the framework of collaborative discourse theory. [ABSTRACT FROM AUTHOR]

The aim of the PPTS project (Pedagogical Platoon Training System) is to design and implement an evaluation environment for strategic and tactical skills, coupled to a network of full-scale simulators of the LECLERC tank. The three-level architecture adopted here is modelled on the levels of expertise, technical, tactic and strategic. The technical level is carried out by reactive agents, the monitors. They point out the important events of the exercise. The evaluation of tactical skills is carried out by cognitive agents, the analysts, which use the facts generated by the monitors to evaluate the tactical skills. The software architecture at analyst level provides a flexible solution to the problem of time in the evaluation of the exercise, whether the transitions between phases are sudden or gradual, whether the exercise is slow or rapid. We will complete this study with the strategic level. [ABSTRACT FROM AUTHOR]

The present paper discusses the results of application of multi-agent technologies to the diagnostic of errors (and underlying cognitive gaps) in aeronautic training. Cognitive psychology issues related to errors constitute the base for ITS modelling. The model of reasoning based on qualitative simulation is proposed as the base for pedagogical evaluation of the trainee. As a result, simulation-based ITS using "learning-by-doing errors" approach is presented, where multi-agent architecture is used for threefold errors diagnosis: ergonomy, knowledge, psychology. The results of this research have found their applications in ASIMIL project. [ABSTRACT FROM AUTHOR]

The know-how transmission becomes an essential condition of high technology companies continuity. In response, we propose a multimedia training platform in operators work station, called SIMPA. It was developed within the framework of a "Six Sigma" step, with a permanent concern of customers satisfaction (learners, formative persons in workshops, and managers). Its foundation is to leave a large place in the autonomous activity of the learner, by insuring him a permanent pedagogical accompaniment, thanks to two computing tutors (the professor and the companion); this answers the need of reduction in time of human tutoring. SIMPA integrates all the tools of follow-up necessary for the trainer to practise an indispensable debriefing at the end of each training session. We experimented SIMPA in situation at TED / TIV; the results are convincing, as well in credibility of our computing tutors as in diversification of pedagogical courses. [ABSTRACT FROM AUTHOR]

This paper describes the implementation of a distance learning flight-training system. We described the various methods used during the implementation of the software (Virtual reality and Case Based Reasoning). ASIMIL proposes a client/server architecture that enables the user to have automated real-time assistance in addition to normal error detection. [ABSTRACT FROM AUTHOR]