Catálogo de publicaciones - libros

In this paper, we present the application of unsupervised learning techniques to automatically recognize behaviors that may be detrimental to learning during interaction with an Exploratory Learning Environment (ELE). First, we describe how we use the -means clustering algorithm for off-line identification of learner groups with distinguishing interaction patterns who also show similar learning improvements with an ELE. We then discuss how a -means on-line classifier, trained with the learner groups detected off-line, can be used for adaptive support in ELEs. We aim to show the value of a data-based approach for recognizing learners as an alternative to knowledge-based approaches that tend to be complex and time-consuming even for domain experts, especially in highly unstructured ELEs.

Item Response Theory (IRT) models were investigated as a tool for student modeling in an intelligent tutoring system (ITS). The models were tested using real data of high school students using the Wayang Outpost, a computer-based tutor for the mathematics portion of the Scholastic Aptitude Test (SAT). A cross-validation framework was developed and three metrics to measure prediction accuracy were compared. The trained models predicted with 72% accuracy whether a student would answer a multiple choice problem correctly.

Atomic Bayesian Networks (ABNs) combine several valuable features in student models: prerequisite relationships, concept to solution step relationships, and real time responsiveness. Recent work addresses some of these features but have not combined them, which we believe is necessary in an ITS that helps students learn in a complex domain, in our case, object-oriented design. A refined representation of prerequisite relationships considers relationships between concepts as explicit knowledge units. Theorems show how to reduce the number of parameters required to a small constant, so that each ABN can guarantee a real time response. We evaluated ABN-based student models with 240 simulated students, investigating their behavior for different types of students and different slip and guess values. Holding slip and guess to equal, small values, ABNs are able to produce accurate diagnostic rates for student knowledge states.

One of the most persistently difficult aspects of vocabulary for foreign language learners is collocation. This paper describes a browser-based agent that assists learners in acquiring collocations in context during their unrestricted Web browsing. The agent overcomes the limitations imposed by learner models in traditional ITS. Its capacity to function in noisy unscripted contexts derives from a well-understood theory of lexical knowledge that attributes a word’s identity to its contextual features. Collocations constitute a central feature type, and we extract these features computationally from a 20-million-word portion of BNC. These we are able to detect and highlight in real time for learners in the noisy Web environments they freely browse. Our learner model, derived by semi-automatic techniques from our 3-million word corpus of learner English, maps detected collocations onto corresponding collocation errors produced by this learner population, alerting learners to the non-substitutability of words within the target collocations. A notebook offers a push function for individualized repeated exposure to examples of these collocations in context.

Each learner has different preferences and needs. Therefore, it is very crucial to provide the different styles of learners with different learning environments that are more preferred and more efficient to them. This paper reports a study of the intelligent learning environment where the learner’s preferences are diagnosed, and then user interfaces are customized in an adaptive manner to accommodate the preferences. A learning system with a specific interface has been devised based on the learning-style model by Felder & Silverman, so that different learner preferences are revealed through user interactions with the system. Using this interface, learning styles are diagnosed from learner behavior patterns on the interface using Decision Tree and Hidden Markov Model approaches.

Motivation plays a key role in learning and teaching, in particular in technology enhanced learning environments. According to motivational theories, proper contingency design is an important prerequisite to motivate learners. In this paper, we demonstrate how confidence levels in an adaptive educational system can be raised using a contingency design technique. Learners that saw parts of a complete picture depending on their performance were more confident to solve the next task than learners who did not. Results suggest that it is possible to raise confidence levels of learners through appropriate contingency design and thus to automatically adapt to their motivational states.

The M-Ecolab was developed to provide motivational scaffolding via an on-screen character whose demeanour defended on modelling the learner’s motivational state at interaction time. Motivational modelling was based on three variables: effort, independence and the confidence. A classroom evalu-ation was conducted to illustrate the effects of motivational scaffolding. Students had an eighty minute interaction with the M-Ecolab, divided into two sessions. The results suggested a positive effect of the motivational scaffolding, particularly for initially de-motivated students who demonstrated higher learning gains. We found that these students followed the suggestions of the on-screen character which delivered personalized feedback. They behaved in a way that was conducive to learning by being challenge-seekers and displaying an inclination to exert more effort. This paper gives a detailed account of the methodology and findings that resulted from the empirical evaluation.

This paper describes the design of the learner modelling component of the system, which was conceived to integrate modelling of learners’ competencies in a subject domain, motivational and affective dispositions and meta-cognition. This goal has been achieved by organising learner models as stacks, with the subject domain as ground layer and competency, motivation, affect and meta-cognition as upper layers. A concept map per layer defines each layer’s elements and internal structure, and beliefs are associated to the applications of elements in upper-layers to elements in lower-layers. Beliefs are represented using belief functions and organised in a network constructed as the composition of all layers’ concept maps, which is used for propagation of evidence.

Self-efficacy is an individual’s belief about her ability to perform well in a given situation. Because self-efficacious students are effective learners, endowing intelligent tutoring systems with the ability to diagnose self-efficacy could lead to improved pedagogy. Self-efficacy is influenced by (and influences) affective state. Thus, physiological data might be used to predict a students’ level of self-efficacy. This paper investigates an inductive approach to automatically constructing models of self-efficacy that can be used at runtime to inform pedagogical decisions. In an empirical study, two families of self-efficacy models were induced: a static model, learned solely from pre-test (non-intrusively collected) data, and a dynamic model, learned from both pre-test data as well as runtime physiological data collected with a biofeedback apparatus. The resulting static model is able to predict students’ real-time levels of self-efficacy with reasonable accuracy, while the physiologically informed dynamic model is even more accurate.

Instant Messaging enables learners and educators to interact in an on-line environment. In this paper, we propose an intelligent ChatBot system, based on instant messaging, for student on-line coaching in an English learning environment. The proposed ChatBot facilitates synchronous communication with students by using ready reference materials including, dictionaries, authorized conversation material with speaking, and a question-answering function. The agent records and analyzes conversations so that the teacher can assess students’ progress. Our contribution in this paper is that we integrate the NLP Tool and AIML into an instant messaging-based ChatBot for English as a Second Language programs.