Chaisak Srisethanil
Graduate Reserch Assistant
chaisak@eiffel.ce.gatech.edu
Nelson Baker, Ph.D
Associate Professor
nelson.baker@ce.gatech.edu
The emergence of computer applications in education and training, such as intelligent tutoring systems, provides a viable alternative to achieve these teaching tasks. The major benefits of the computerized instruction are several. First, the computer is never exhausted with a student if they request multiple repetition of instruction. Second, without the nature of human mistake, the computer always provides instruction that adapt to learning condition. Thus computerized instruction is a significant tool for the effective application of adaptive teaching.
ITS-Engineering is a tutoring system shell intended to provide a developing framework for applications in the engineering domains with less time and cost. Based on Gagne's instructional design and the multiple teaching style paradigm, the application is able to deliver instruction that adapts in both content and teaching styles. The available teaching styles in ITS-Engineering include instructor-oriented, guided-discovery, user-initiated and exploratory styles. The instructor-oriented and the guided discovery style represents the teacher-control paradigm. In contrast, the user-initiated and the exploratory style represent the learner-control paradigm. The application of these teaching styles and their adapting capabilities are demonstrated in ITS-CPM (Intelligent Tutoring System for Construction and Project Management); an ITS application developed with in the framework of ITS-Engineering.
The emerging roles of computer technology in the education and training community provide a plausible platform to deliver adaptive instruction. Adaptive teaching, adaptive education, and adaptive instruction are fairly new terms that are frequently used to describe individualized approaches to education and training. Adaptive teaching may be defined as optimal instruction that effectively meets the individual needs of the students and is conveyed toward a student's different aptitude i.e., intellectual abilities, personalities, and cognitive styles of learning [5]. The goals of adaptive teaching are the implementation of diagnostic and prescriptive instructional systems designed to make learning more effective, efficient and meaningful.
Adaptive teaching can be implemented at either the macro level or the micro level [2]. The macroadaptation happens in a longer period of time (e.g., month-to-month) while the microadaptation occurs moment-to-moment as seen in several computerized instructional programs such as ITS-Engineering.
ITS-Engineering is a research program within the project EPITOME (Engineering Platform for Intelligent TutOrs and Multimedia Experiences) [1]. The main objectives of ITS-Engineering are to provide the developing framework for applications in the engineering domains with less cost and time. Additionally, an application developed within ITS-Engineering framework is able to deliver instructions that are adaptive in both topic contents and teaching styles. The following section describes the adaptive capability of the system followed by the implementation of ITS-Engineering and its domain-dependent application, ITS-CPM.
The term adaptive teaching in ITS-Engineering includes the instructions that are not only adapting the contents of the domain to match the level of knowledge of a student, but also adapting the styles of teaching to match the changing learning/teaching conditions. ITS-Engineering allows the instruction to adapt its content by maintaining a student model [7]. The instruction is adapted to the level of the student knowledge reflected in the student model.
Additional to the content adaptive instruction, ITS-Engineering takes another step by providing the instructions in different styles of teaching. Combined with the content adaptive teaching methodology, the adaptive multiple teaching styles of ITS-Engineering thoroughly address the issue of how and when to teach what. The instructions of the systems are designed and implemented based upon Gagné's instructional design framework [3] that is widely accepted and independent of subject domains. The instructional framework primarily consists of the desired learning outcomes (e.g., rules, problem solving) and the set of instructional events and the attached learning activities to achieve a certain outcome. The sequence, content and number of instructional events vary upon the learning outcomes; thus resulting in multiple teaching styles. The set of instructional events is the building block for developing multiple teaching style instruction in ITS-Engineering.
ITS-Engineering's teaching styles include the instructor-oriented and the guided-discovery style representing the teacher-control teaching style and the user-initiated and the exploratory style representing the learner-control approach [8].
Instructor-Oriented (I-0) The key characteristic of this style is its directness and control. The system directly presents and demonstrates the knowledge to the student; and selects what the student should learn from one moment to the next. The style is considered efficient in regard to learning and teaching time [7]
Guided Discovery (G-D) Instead of directly presenting the information to the student as characterized in the I-O style, the guided discovery (as the name implies) indirectly presents the information by asking the student questions that intrigue them to think, to reason and to discover the concepts.
User-Initiated (U-I) The rationales of the U-I style are practicing and learning on demand. With the style, a student can work on an exercise that is dynamically generated customized to the student's needs. The U-I style allows the student to practice on the problem she wants to emphasize by creating her own exercise based on different parameters. This style is very appropriate to most engineering domains that involve mathematical computation and problem solving.
Exploratory (Ex) In the exploratory style, a student is free to choose which topics to learn. The learner fully directs the path and pace of the learning session. The Ex style enables a student to overview the content of the subjects. Consequently, the student can develop an expectation for a particular subject that is beneficial for subsequent instruction.
Figure 1 displays (with the emphasis on instructor module) the architecture of ITS-Engineering consisting of four main components: a student module; a domain module; an authoring tool; and an instructor module.
The student module represents the student's present level of domain knowledge; thus enabling the system to adapt the content of instruction to match that level. The domain module contains the high level knowledge and the low level knowledge. The high level contains the knowledge element that is independent of subjects (e.g., prerequisite, type of instructional element, media type, etc.). The lower level includes the topic specific information (e.g., facts, procedure, concepts, etc.). The authoring tool contains instructional design knowledge to provide consultation to an instructor in designing instruction, collecting domain knowledge, and selecting teaching styles. The development of the authoring tool is ongoing and not yet fully implemented.
The instructor module contains the domain independent pedagogical knowledge, (e.g., teaching styles). Since this paper focuses on describing adaptive instruction in the computer tutor, the instructor module and its mechanisms is subsequently described.
The instructor module performs two major functions: delivery of instruction across subject domains; and delivery in different teaching styles. The instructor module consists of the instructor model, the instruction generator, and the pedagogical knowledge base as seen in Figure 1. The instructor model determines what to teach, the instruction generator details how to teach, and the pedagogical knowledge base contains strategies of when to teach and in which style.
The instructor model controls the sequence and content of an instructional event during a tutoring session via the curricula/discourse manager. These instructional events vary upon teaching style. For instance, the set of instructional events contains present information, elicit performance, and feedback in the U-I style. The evaluation/response unit assesses the student's performance and provides feedback to the student. The type of evaluation and response vary upon the different teaching styles.
The main task of the Instruction generator is to create and deliver instruction in different teaching styles and across various domains. The generator creates instruction in two different modes: static and dynamic. In static mode the generator simply retrieves the matched information from the domain module and then displays it to the learner. When the generator needs to dynamically generate the instruction (e.g., in the U-I style), it uses the matched prototype problem to produce the novel problem with different generated parameter values [6].
The pedagogical knowledge base, i.e., how &when to teach what, contains three major categories of the information: the backbone instructional rules based on Gange's instructional design, the knowledge of each teaching style and the meta knowledge of teaching styles [6]. Together, this knowledge provides the theoretical framework on how to generate and deliver adaptive multiple teaching styles.
Based on the framework of ITS-Engineering, two applications have been developed: ITS-CPM and ITS-Shear [6], [7]. ITS-CPM has been developed to teach construction planning and scheduling techniques. The application contains subjects of general planning concepts, CPM, precedence diagram, PERT, Line of Balance, crashing and time-cost trade-offs and resource allocation.
ITS-Shear is developed for tutoring the analysis of shear failure in the reinforced concrete structure. It covers topics in calculating shear of beams with and without stirrups, rectangular beams and the T-beams. In contrast to the diverse domain of ITS-CPM, i.e., declarative and procedural, ITS-Shear is mostly procedural.
Both applications are in engineering domains but significantly different in nature. However, ITS-Engineering's framework is sufficient to facilitate the development of these applications. The plausibility of multiple teaching styles to engineering education is not fully investigated and is being evaluated. However, the literature indicates the advantage of the paradigm to other applications and we are convinced that the strategy is beneficial to engineering subjects, as well. Additionally, ITS-CPM and ITS-Shear illustrate the potential mechanism to provide assistance to classroom instruction and a source for continuing education in the engineering community.
The following descriptions of learning sessions are based on an application-ITS-CPM developed within the framework as described above. There are three learning sessions in ITS-Engineering: tutoring, practicing and browsing (see Figure 2). A student is free to enter the session she desires.
