Claude Frasson

Analyzing a new learning strategy according to different knowledge levels

Abstract Intelligent tutoring systems have recently evolved towards a co-operative approach between the learner and the system. Knowledge acquisition is facilitated by interaction with the system under the control of the learner. New tutoring strategies have been introduced to enhance motivation of the learner by involving a second learner or a companion who simulates the behaviour of a second learner in the learning process. An inverted model called “learning by teaching” in which the learner could teach the learning companion by giving explanations has also been presented. In this paper we discuss the advantage and the inconvenience of these strategies and present a new learning strategy which improves performance for good or intermediate learners. We describe an experiment with this strategy and compare results with those obtained with the companion. We analyze and discuss results obtained.

An Actor Based Architecture for Intelligent Tutoring Systems

The evolution of intelligent tutoring systems (ITS) toward the use of multiple learning strategies calls on a multi-agent architecture. We designed an ITS where several agents assume different pedagogical roles; consequently, we called them actors. We first describe the conceptual architecture of an actor which allows it to be reactive, instructable, adaptive and cognitive. We then provide a detailed view of this architecture and show how it functions with an example involving the different actors of a new learning strategy, the learning by disturbing strategy.

An Integrated Approach for Automatic Aggregation of Learning Knowledge Objects

This paper presents the Knowledge Puzzle, an ontology-based platform designed to facilitate domain knowledge acquisition from textual documents for knowledge-based systems. First, the Knowledge Puzzle Platform performs an automatic generation of a domain ontology from documents’ content through natural language processing and machine learning technologies. Second, it employs a new content model, the Knowledge Puzzle Content Model, which aims to model learning material from annotated content. Annotations are performed semi-automatically based on IBM’s Unstructured Information Management Architecture and are stored in an Organizational memory (OM) as knowledge fragments. The organizational memory is used as a knowledge base for a training environment (an Intelligent Tutoring System or an e-Learning environment). The main objective of these annotations is to enable the automatic aggregation of Learning Knowledge Objects (LKOs) guided by instructional strategies, which are provided through SWRL rules. Finally, a methodology is proposed to generate SCORM-compliant learning objects from these LKOs.

Using machine learning to predict learner emotional state from brainwaves

Intelligent Tutoring Systems (ITS) learner model has progressively evolved. Initially composed of a cognitive module it was extended with a psychological module and an emotional module. The learner model still remains non-exhaustive. Methods of data collection on the cognitive and emotional state of the learner often lack precision and objectivity. In this paper we introduce an emomental agent. It interacts with an ITS to communicate the emotional state of the learner based upon his mental state. The mental state is obtained from the learners brainwaves. The agent learns to predict the learners emotions by using machine learning techniques.

LANCA: A Distance Learning Architecture Based on Networked Cognitive Agents

The use of Internet as a general vehicle to support distance learning is a recent orientation of learning with multiple positive and negative consequences. The important disadvantage of such an approach is to forget the difference between information and knowledge, between consultation and pedagogy, leading to poor training as a consequence. The need of providing access to information to a larger number of people should not be realized to the detriment of the quality of training. In this paper we first explain why and how ITS techniques using intelligent agents can be adapted to distance learning. We precise the main characteristics of these agents and their functions in a distributed environment. We then present the architecture of this environment with the role of the different intelligent agents. We show, on an example, how the agents interact with the learner and particularly how a pedagogical agent can switch to a new strategy according to the progression of the learner.

Cooperative learning strategies for intelligent tutoring systems

Intelligent tutoring systems (ITS) are evolving towards a more cooperative relationship between the system and the student. More and more, learning is considered as a constructive process rather than a simple transfer of knowledge. This trend has brought to light new cooperative tutoring strategies. One of these tutoring strategies, the learning companion, designed to overcome some of the limitations of the classical tutoring model, involves a student and two simulated participants: a tutor and another student. More recently, a new strategy, learning by disturbing, has been proposed. In this strategy, the simulated student is a troublemaker whose role is to deliberately disturb the human student. This article describes the learning by disturbing strategy by contrasting it with the learning companion strategy. In addition, links are drawn between this new strategy and the psychology of learning, in particular the cognitive dissonance theory. An indicator has been developed that measures discord between the ideas, helping to pinpoint the concepts that are most likely to be misunderstood by the learner. Doing so allows one to plan more efficiently the interventions of the troublemaker.

CREAM-Tools: An Authoring Environment for Knowledge Engineering in Intelligent Tutoring Systems

This chapter presents an authoring model and a system for curriculum development in Intelligent Tutoring Systems (ITSs). We first present an approach for modeling knowledge of the subject matter (the curriculum) to be taught by a large-scale ITS, and we show how it serves as the framework of the authoring process. This approach, called CREAM (Curriculum REspresentation and Acquisition Model), allows creation and organization of the curriculum according to three models concerning respectively the domain, the pedagogy and the didactic aspects. The domain is supported by the capability model (CREAM-C) which represents and organizes domain knowledge through logical links. The pedagogical view allows the definition and organization of teaching objectives by modeling skills required to achieve them and evaluating the impact of this achievement on the domain knowledge (CREAM-O and pedagogical model). The didactic component is based on a model of resources which defines and specifies different activities that are necessary to support teaching (CREAM-R). The construction of each part of CREAM is supported by specific authoring tools and methods. The overall authoring system, called CREAM-Tools, allows Instructional Designers (IDs) to produce a complete ITS curriculum based on the CREAM approach. Although this article is limited to curriculum development, we give some guidelines on how the resulting system could support the construction of other ITS components such as the planner and the student model.

Modeling mental workload using EEG features for intelligent systems

Endowing systems with abilities to assess a users mental state in an operational environment could be useful to improve communication and interaction methods. In this work we seek to model user mental workload using spectral features extracted from electroencephalography (EEG) data. In particular, data were gathered from 17 participants who performed different cognitive tasks. We also explore the application of our model in a non laboratory context by analyzing the behavior of our model in an educational context. Our findings have implications for intelligent tutoring systems seeking to continuously assess and adapt to a learners state.

Subliminal cues while teaching: HCI technique for enhanced learning

This paper presents results from an empirical study conducted with a subliminal teaching technique aimed at enhancing learners performance in Intelligent Systems through the use of physiological sensors. This technique uses carefully designed subliminal cues (positive) and miscues (negative) and projects them under the learners perceptual visual threshold. A positive cue, called answer cue, is a hint aiming to enhance the learners inductive reasoning abilities and projected in a way to help them figure out the solution faster but more importantly better. A negative cue, called miscue, is also used and aims at obviously at the opposite (distract the learner or lead them to the wrong conclusion). The latest obtained results showed that only subliminal cues, not miscues, could significantly increase learner performance and intuition in a logic-based problem-solving task. Nonintrusive physiological sensors (EEG for recording brainwaves, blood volume pressure to compute heart rate and skin response to record skin conductivity) were used to record affective and cerebral responses throughout the experiment. The descriptive analysis, combined with the physiological data, provides compelling evidence for the positive impact of answer cues on reasoning and intuitive decision making in a logic-based problem-solving paradigm.

CREAM-Tools: An Authoring Environment for Curriculum and Course Biulding in an Intelligent Tutoring System

The main goal of the third stage of the SAFARI project is the delivery by an intelligent tutoring system (ITS) for a complete course. To achieve this goal, a subject-matter model, called CREAM, that can support course generation and delivery has been proposed. The acquisition of such a knowledge model requires to enable the designer with dedicated tools and methods. The purpose of this paper is to present the authoring environment (CREAM-Tools) we developped to support course and curriculum construction using the CREAM approach. This authoring environment consists of a course generation kit, of building tools and methodologies. Curriculums and courses produced with this environment are directly usable by an ITS. We also show ways other modules in an intelligent tutorial system can exploit the resulting curriculums and courses.