Showing papers on "Intelligent tutoring system published in 1992"

Computer-assisted instruction and intelligent tutoring systems : shared goals and complementary approaches

Abstract: Contents: JH Larkin, RW Chabay, Introduction S Dugdale, The Design of Computer-Based Mathematics Instruction GR Culley, From Syntax to Semantics in Foreign Language CAI AT Corbett, JR Anderson, LISP Intelligent Tutoring System: Research in Skill Acquisition BJ Reiser, DY Kimberg, MC Lovett, M Ranney, Knowledge Representation and Explanation in GIL, An Intelligent Tutor for Programming RW Chabay, BA Sherwood, A Practical Guide for the Creation of Educational Software G Brackett, Realizing the Revolution: A Brief Case Study A Lesgold, S Lajoie, M Bunzo, G Eggan, SHERLOCK: A Coached Practice Environment for an Electronics Troubleshooting Job W Sack, E Soloway, From PROUST to CHIRON

A Framework for Intelligent Knowledge Sequencing and Task Sequencing

Abstract: Most effective human tutors possess the skill of adaptive sequencing of knowledge and tasks. This skill is also the key function of many important tutoring systems and learning environments. A number of workers in the field of intelligent tutoring systems have tried to build a framework for intelligent knowledge and task sequencing. In this paper we briefly discuss previous work on building a framework and strategies for knowledge and task sequencing. Then we suggest several additional components we have designed to complete the framework and describe a pragmatic strategy for multiple-kind, multiple-concept task sequencing based upon the framework.

Probabilistic Student Models: Bayesian Belief Networks and Knowledge Space Theory

Abstract: The applicability of Knowledge Space Theory (Falmagne and Doignon) and Bayesian Belief Networks (Pearl) as probabilistic student models imbedded in an Intelligent Tutoring System is examined. Student modeling issues such as knowledge representation, adaptive assessment, curriculum advancement, and student feedback are addressed. Several factors contribute to uncertainty in student modeling such as careless errors and lucky guesses, learning and forgetting, and unanticipated student response patterns. However, a probabilistic student model can represent uncertainty regarding the estimate of the student's knowledge and can be tested using empirical student data and established statistical techniques.

A Comparison of Learning Environments: All That Glitters

Abstract: : Aptitude-treatment interactions (ATI) refer to the covariation between learner characteristic and instructional treatment in relation to some outcome measure To systematically test for ATI, I used an intelligent tutoring system instructing basic principles of electricity as a complex but controlled learning task I created two instructional environments from this one tutor, differing only in feedback In the rule-application environment, the system provided learners with relevant principles, and in the rule-induction environment, learners had to induce principles on their own The learner characteristic examined in this paper was 'exploratory behavior,' a measure of on-line tool usage I hypothesized that exploratory learners would learn faster and better if they had been assigned to the inductive environment and less exploratory learners would benefit from the more structured, application environment Results showed significant aptitude-treatment interactions in the data, confirming the above hypotheses Implications of these findings are discussed in relation to the design of intelligent tutoring systems

The Search for Adaptability, Flexibility, and Individualization: Approaches to Curriculum in Intelligent Tutoring Systems

Abstract: Curriculum is a pervasive concern in education, but is barely an issue in intelligent tutoring system (ITS) research. This is because most intelligent tutoring systems are restricted to working in a tightly constrained domain, and thus have little need for subject organization at the macro level of curriculum. Nevertheless, there is a small, but interesting, line of research in ITS that has explored issues of direct concern to curriculum. This research has led to the idea of curriculum as an emergent phenomenon, created dynamically in response to student needs, to requirements of the subject matter, and to pedagogical goals. Such a curriculum is adaptable to a changing environment, flexible in instructional goals, and individualized to the student. This paper traces the evolution of the notion of curriculum in intelligent tutoring systems, in order to show how and why the current idea of curriculum came about, and to demonstrate its power. Along the way, particular systems are used to illustrate interesting aspects of ITS curriculum and to give concrete foundation to the discussion.

Conceptual Modelling in Error Analysis in Computer-Assisted Language Learning Systems

Abstract: Many computer-assisted language learning systems specifically designed to be used in the curriculum and which exploit AI techniques have neither a learner model nor consequently any deep error analysis. Evidence from applied linguistics shows that learners have their own system of rules for the production of a foreign language. We believe the central issue is to determine the appropriate level of description of these rules and uncover the strategies used by the learners in particular situations. This information represents the major part of the learner model. We review error analysis in second language learning and tutoring systems related to this perspective. We introduce a new structure, called an “applicable rule”, that can be used to help diagnose and to represent a learner’s performance. We propose a design for the architecture of a system for computer diagnoses of learners’ grammatical performances in a communicative environment. Examples of diagnosis using applicable rules illustrate the functioning of this architecture.

COCA: A Shell for Intelligent Tutoring Systems

Abstract: There are several reasons why intelligent tutoring systems (ITSs) have failed to gain widespread acceptance in the classroom. These include cost (ITSs often run on platforms that are too expensive for schools). Also, many ITSs are restricted to one particular domain and do not allow teachers to configure them for other domains. From interviews with teachers we identified yet a further reason: most ITSs teach according to a fixed teaching strategy, and do not allow teachers to alter the way in which material is taught. In this paper, we describe a system that allows one to do so. The system, COCA (CO-operative Classroom Assistant), contains a number of user-changeable control heuristics which implement decisions that need to be made during teaching.

Intelligent tutoring system for electric circuit exercising

Abstract: An intelligent tutoring system called Circuit Exerciser is described. The system is designed to help university students learn more about electric circuits. It can formulate drill problems, solve them, and infer mistakes in a student's answer. It can also provide helpful comments to the students on how the mistake was made. The system shows the circuit of the presented problem on graphic displays and is student-friendly. The system architecture, pedagogical cycle, and execution of the system are presented. >

Student Strategies for Learning Programming from a Computational Environment

Abstract: This paper discusses the design and evaluation of a hypertext-based environment that presents instructional material on programming in Lisp. The design of the environment was motivated by results from studies investigating students' strategies for knowledge acquisition. The effectiveness of the design was evaluated by conducting a study that contrasted how subjects used and learned from the instructional environment compared to subjects using more standard, structured, linear instruction. The results showed an interesting ability by environment interaction: the higher ability subjects using the hypertext environment improved and made significantly less errors when programming new concepts while the lower ability subjects did not improve and made more errors. Meanwhile, subjects using the control environment did not show this ability-based difference. These results have implications for the design of intelligent tutoring systems. They affect decisions involving the amount of learner control that is provided to students and the way student models are constructed.

Distributed Learning Companion System: WEST Revisited

Abstract: This paper describes a distributed learning system which consists of two connected computers so that students can learn in collaboration and/or competition at different locations. Considering different numbers and roles of involved agents, we have enumerated 768 possible distributed learning models. Among them, we evaluated 3 models using a system which is a reimplementation of the well-known WEST program. The evaluation result has two significant implications: (1) such learning systems hold the promise to be a form of futuristic intelligent computer classroom and (2) competition could be a powerful motive in learning that would shed new light on the Intelligent Tutoring System research.

Instructional planning in an intelligent tutoring system: combining global lesson plans with local discourse control

Abstract: This thesis focuses on the design and development of an instructional planner for an intelligent tutoring system for Cardiovascular Physiology, that will assist first year medical students to learn the causal relationships between the parameters of the circulatory system and to solve problems about simulated disturbances to the system. The instructional planner is responsible for selecting or generating instructional goals, deciding how to teach the selected goals, monitoring and critiquing the student's behavior, and determining what to do next at each point during a tutoring session. The approach taken in this thesis is to build a planner that integrates opportunistic control with sophisticated instructional planning methods: combining capabilities of lesson planning with discourse planning. The lesson planning is further divided into goal generation, planning of strategies, and planning of tactics to refine the goal into subgoals. The discourse planning is implemented using a two level approach: pedagogical decision making at the upper level and tactical discourse state-based planning at the lower level in its discourse management network. The planner plans dynamically based on the inferred student model; it generates plans, monitors the execution of the plans, and replans when the student interrupts with a question during the tutoring session. The planner supports a mixed-initiative strategy, allowing student initiatives during the tutoring session. The planner needs to do replanning after it carries out the student's request. The pedagogical knowledge is extracted from the experts and represented explicitly as rules, lesson planning rules and discourse planning rules. The system interprets the rules and builds the lesson plans or returns an appropriate discourse action. The complete set of tutoring rules, a short trace of a tutoring session that describes the operation of the system at every step, and sample dialogues produced by the system can be found in the appendices.

Computer dialogue system (CDS): A system for capturing computer-mediated dialogue

Abstract: In the process of studying human tutoring sessions as a basis for building an intelligent tutoring system, we developed a computer dialogue system (CDS) that allows two PC users to communicate with each other over a telephone line by typing at a computer keyboard. CDS records the content of dialogue on a disk in a specified, well-formatted way. It also makes available a way to mimic some of the characteristics of face-to-face dialogue such as repair. It was developed in the Smartcom III (Hayes Communication package) environment. Thus, it is fast, portable, and easy to use. In addition, to help us study the recorded dialogues, we have also written a numbering program to label each turn and each sentence within each turn. Although CDS was originally designed for the study of tutorial dialogues between students and teachers, it can be used to conduct and record dialogues of any kind.

Automated Generation of Examples for a Tutorial in Case-Based Argumentation

Abstract: Examples in our domain are argumentation problems involving small sets of legal cases that are related in interesting ways We call these collections of cases Argument Contexts We have identified several types of Argument Contexts that can be used to teach the Issues in our curriculum for case-based argumentation We have developed a program that generates Argument Contexts that address these Issues and have other pedagogically desirable properties, such as being clear and concise In a preliminary feasibility study a human tutor used examples generated by the program with encouraging results Because assembling Argument Contexts by hand is time-consuming, we believe that our program can be a useful tool for law professors who teach reasoning with cases Also, we plan to use it as a component of an intelligent tutoring system for case-based argumentation that we are currently developing

Knowledge Representation for an Intelligent Tutoring Sytem Based on a Multilevel Causal Model

Abstract: An explicit representation of knowledge is central for an Intelligent Tutoring System (ITS). In order for a system to acquire the necessary flexibility, its knowledge representation framework should distinguish between several types of knowledge and structure them in layers. Here, we present a method for representing domain knowledge for an ITS, using hierarchical knowledge structures and a multilevel causal model of the domain. The successive levels of this causal model increase in complexity to more closely approximate a complete domain model. The resulting knowledge structures have the flexibility that is needed to invoke a sophisticated instructional session.

Results of encoding knowledge with tutor construction tools

Abstract: We have developed and evaluated a set of tutor construction tools which enabled three computer-naive educators to build, test and modify an intelligent tutoring system The tools constitute a knowledge acquisition interface for representing and rapid prototyping both domain and tutoring knowledge A formative evaluation is described which lasted nearly two years and involved 20 students This research aims to understand and support the knowledge acquisition process in education and to facilitate browsing and modification of knowledge Results of a person-hour analysis of throughput factors are provided along with knowledge representation and engineering issues for developing knowledge acquisition interfaces in education

The Grace Tutor: A Qualified Success

Abstract: The Grace Intelligent Tutoring System is being developed within NYNEX Science & Technology, Inc. to assist in teaching COBOL to both novice and experienced programmers. We describe the result of three trials with Grace and the lessons that we learned by having a wide variety of student populations in real classes. We discuss the changes to the Grace Tutor which create greater congruence between the tutor and the actual programming environments used by the students. Finally we discuss the one aspect in which we do not consider the Grace Tutor a success: our inability to move Grace from the lab into everyday, classroom use. We discuss why we think this critical point has not yet been reached, not only by the Grace Tutor but by numerous other educational technology projects which have been only laboratory successes.

The Uses of Multiple Student Inputs in Modeling and Lesson Planning in CAI and ICAI Programs

Abstract: Responding appropriately to student errors requires some model of the student with which to determine the most likely cause of the errors. In a conventional CAI program the model is implicit and is represented by the hard-coded relationship between errors and corrective feedback. In an intelligent tutoring system (ICAI program) student modeling can be done dynamically as student responses are generated. In both cases, multiple inputs about causally related variables obtained prior to any tutoring provides a rich source of information about the cognitive state of the student. As a result it is possible to produce a more robust student model and to generate a more effective sequence of lessons to repair the student's misconceptions. Examples of such an approach used in the implementation of both a CAI and a ICAI program are presented.

Curriculum Tree: A Knowledge-Based Architecture for Intelligent Tutoring Systems

Abstract: This paper describes a knowledge-based architecture, called curriculum tree, for building intelligent tutoring systems. Primarily based on the subject domain knowledge structure, the architecture naturally incorporates the global curriculum planning and monitors the local learning activities. The curriculum tree can also be viewed as a structure of various teaching knowledge at different stages of learning. By adopting rule inheritance, the architecture allows additional additivity and flexibility for developing an intelligent tutoring system incrementally as well as efficiency for running rules in each learning episode. Thus, curriculum tree is an architecture towards building large scale intelligent tutoring systems. In this paper, we shall also discuss how the curriculum tree architecture is used in building Integration-Kid, a Learning Companion System which is a particularly complex type of intelligent tutoring system, in the domain of learning indefinite integration.

Using tutoring systems to study learning: An application of HyperCard

Abstract: HyperCard was used to develop a simplified tutoring system whose principles were based on a learning theory, and a genetics tutoring system was evaluated experimentally. Learning was studied by examining immediate versus delayed feedback after an error was made. Such tutoring systems aid in psychological studies of learning, because experimental variables can be easily manipulated. HyperCard provides a good vehicle for tutoring system development, since it requires no extensive programming skills.

QUIZ, a Distributed Intelligent Tutoring System

Abstract: QUIZ is a Distributed Intelligent Tutoring System for learning bridge bidding. A set of generic tasks is distributed among four specialists (a tutor, a problem solver, an explainer and a problem generator), who perform their tasks in parallel. These agents are heterogeneous. They own a private working memory and communicate by asynchronous messages passing. Increasing flexibility is a key point in improving the pedagogical capacities of ITS. The strategic level of flexibility results from the determination of the curriculum, the choice of the pedagogical strategy and the degree of expertise used in the problem solver. The choices made at this level must be based mainly upon the student model. The tactical level of flexibility results from the choice of exercise, the advising, the corrections and the explanations. In QUIZ, the pedagogical actions are gathered and sequenced by means of plans, which are chosen by metarules and dynamically assembled from pieces that are memorized in libraries.

Naive vs. Formal Grammars: A Case for Integration in the Design of a Foreign Language Tutor

Abstract: The relationship between formal and naive (i.e., used for didactic purposes) grammars in foreign language teaching is dealt with in the paper which presents, as a case study, an attempt to integrate the two approaches within an intelligent tutoring system. This work has been carried on in the framework of the ET (English Tutor) project whose long term goal is the development of a tutoring system aimed at helping Italian students master English verb tenses. Within the project, a prototype system based on a naive approach to the grammar of tense has been built. The experimentation performed with the prototype provided the motivation for a critical re-evaluation and revision of some of the assumptions which it was grounded upon. The possibility of integrating some naive intuitions into a systemic representation of grammatical knowledge is discussed in the paper, and a new version of the domain expert module exploiting the systemic approach to tense selection is illustrated.

Developing the didactic operations for intelligent tutoring systems: A synthesis of artificial intelligence and hypertext.

Abstract: This thesis is concerned with Intelligent Tutoring Systems. It investigates the architecture of an Intelligent Knowledge Based Tutoring System in terms of three knowledge models: that of the domain, the student and the tutor, and examines the interrelatedness and interconnectedness of these three knowledge models. Existing Knowledge Based Tutoring Systems are reviewed, and the relationship between their behaviour and architecture is analysed by evaluating them against Wenger's model of a didactic operation. Two such systems, PROUST, a tutoring system for Pascal program debugging skills, and micro-SEARCH, a tutoring system for mathematical transformations skills, are used in the study. This evaluation serves two purposes: to unravel the requirements for interrelatedness and interconnectedness between the three knowledge models in order to develop a true Knowledge Based Tutoring System with a full-scale didactic operation, and to uncover the limitations of the current generation of Knowledge Based Tutoring Systems and how they fail to fully encompass these requirements. On this basis the thesis goes on to propose a hybrid model made up of Artificial Intelligence and Hypertext concepts which helps to overcome the limitations of existing Knowledge Based Tutoring Systems. This model in particular addresses the requirements for the development of an Intelligent Tutoring Systems with a full-scale didactic operation. The model integrates Hypertext's explicit information nodes and linking properties with Artificial Intelligence's logical inferencing on knowledge representation schemes. The thesis finally shows how to use this model to design a generic Intelligent Tutoring System that supports a full-scale didactic operation.

Building an Intelligent Second Language Tutoring System from Whatever Bits you Happen to Have Lying Around

Abstract: The Computing Research Laboratory (CRL) at New Mexico State University is currently engaged in the design of language teaching software, based on previously developed mature artificial intelligence and machine translation technologies within CRL. Our approach is unique because it uses the robustness of a natural language processing (NLP) system which incorporates both general world knowledge and task domain knowledge (Metallel), beliefs ascription (ViewGen), and semantic parsing techniques (PREMO) in the service of better student-system interaction.

A task-driven approach to knowledge acquisition, analysis and representation for intelligent training systems

Abstract: The authors present work in progress to develop a task driven approach to acquiring, analyzing and representing knowledge for an intelligent tutoring system (ITS) for training operators and maintainers of a process control system. The research premise is that using a task model approach provides direction and infrastructure for acquiring, analyzing, and representing domain-specific knowledge and makes possible the development of an ITS that teaches domain users to perform tasks in the right way at the right time. To demonstrate the efficacy of a task-based approach to knowledge analysis and representation, the authors contrast it with two alternative approaches: component-based and problem-based approaches to knowledge analysis and representation. The task-based approach described not only contains elements of both component- and problem-based approaches, but does so within the context of the specific job being performed, and the function of the person performing that job. >

Grace meets the “real world”: tutoring COBOL as a second language

Abstract: Grace is an intelligent tutoring system for COBOL which has been used to teach both novice and experienced programmers. While the tutor was quite effective in several classes and was designed with cognitive and interface principles in mind, we discuss a number of interesting issues that we have discovered when novice and experienced programmers used the tutor. Most of these problems are related to incompatibilities between the tutor interactions and the students' expectations in two areas: (1) the interactions with the tutor versus the interactions in their usual work environment and (2) the way in which experienced programmers solve problems. We describe these issues along with our solutions in the revised version of the tutor.

Phenomenographic Research and the Design of Diagnostic Strategies for Adaptive Tutoring Systems

Abstract: Adaptive tutoring systems require explicit information about the forms of misconception students have within the topic being taught. Phenomenographic research, which uses qualitative interview data to describe the “outcome space” for learning a specific topic, can provide that level of detail needed. The paper gives examples of this and argues that this approach is more tractable than student modelling and leads to a system architecture that combines knowledge of the student with the interactive interface to define a diagnostic strategy for an adaptive, though not intelligent tutoring system.

Intelligent Tutoring System for Symbolic Calculation

Abstract: In this paper, an ITS-DIITS, which helps student to learn calculus, is introduced as a prototype The whole tutoring procedure of DIITS is divided into three phases: basic, advanced and exploratory exercise In different phrases, the system provides problems embodying skills or strategies of different levels and allows operations of corresponding levels for student to use When the student solves a problem step by step on the exercise interface, the system corrects his errors in intermediate steps in basic exercise phase, or makes advice and comments on his problem solving strategies in advanced phase The domain knowledge, which is represented in the form of an AND/OR graph of skills, directs the search process of the problem solver, and provides template for student modeling and reference for problem generation

The Nihongo Tutorial System: An Intelligent Tutoring System for Technical Japanese Language Instruction.

Abstract: The Nihongo Tutorial System is an intelligent tutoring system designed to assist English-speaking scientists and engineers in acquiring reading proficiency in Japanese technical literature. The system provides individualized lessons by matching a student's technical area of interest and Japanese language ability with the available instructional materials which are derived from actual technical articles. This approach is designed to maximize comprehension through context by assuring an appropriate amount of new material and by drawing on the student's expertise in the subject matter. The instructional texts are encoded in a data structure that maintains syntactic, semantic, phonetic, and morphological information that can be delivered to the student upon request. The system also provides on-line supplementary materials including a character dictionary, a general dictionary, a grammar dictionary, and several technical dictionaries for additional information. Thus the Nihongo Tutorial System provides an efficient aid for acquiring the vocabulary, syntax, and style of Japanese texts from a specific technical discipline.