ITSs can be defined as computerized systems used for tutoring and having the following characteristics (Anohina, 2007): (1) they use principles and methods from artificial intelligence (like, knowledge representation, reasoning, natural language processing and machine learning); (2) are adaptive systems that adapt the tutoring process to the characteristics of the learner, so carrying out adaptive or individualized tutoring; (3) simulate human tutor; (4) are based on the cognitive theory. One of the main characteristics of ITSs is that they try to simulate human tutor to implement adaptive tutoring. To cover all activities done by the human tutor the system has to do the following tasks: generate curriculum, provide learning materials and problems for the learner to solve in each topic, evaluate learner’s knowledge and give meaningful feedback to the learner to help him/her improve his/her knowledge (Lavendelis, 2009). In adaptive tutoring these activities must be carried out adaptively – each learner should receive individual approach that fits his/her characteristics and/or preferences. Curriculum should be generated to meet the needs of individual learner as well as materials and problems should be adapted to each learner.
Simulation of the human tutor is a complex task for the system. It requires intelligent choices and actions to be made. All actions by the teacher, like creation of the curriculum, choice of the appropriate learning materials or evaluation of the learner’s knowledge are complex actions and require intelligence. As a consequence, various intelligent mechanisms are needed for an ITS to simulate such actions. Intelligent mechanisms used to implement adaptive tutoring vary from system to system. Still, the main types of knowledge used in different ITSs are the same. Knowledge about the domain or subject is needed to know what to teach. Knowledge about the learner is needed to know whom to adapt and knowledge about the tutoring process is needed to know how to teach and how to adapt to the learner’s characteristics. The goal of ITSs is to use the above-mentioned three types of knowledge to carry out adaptive tutoring. Usually each type of the knowledge is used by different intelligent mechanisms. It is beneficial to define components corresponding to the three main types of knowledge used in ITSs. As a consequence traditional modular ITS architecture consists of three modules, namely, the expert module, the student diagnosis module and the tutoring module, all together named traditional trinity (Grundspenkis & Anohina, 2005). Additionally the communication module is added to manage the user interface, resulting in the modular architecture consisting of 4 modules. The modular architecture is widely used in intelligent tutoring systems, for example, in Ines (Hospers et al, 2003), AlgeBrain (Alpert et al, 1999), FLUTE (Devedzic et al, 2000) and IKAS (Vilkelis et al, 2009) systems. Modules have the following features (Grundspenkis & Anohina, 2005):
- The expert module represents the domain expert’s knowledge and includes problem solving characteristics. The task of the module is to solve domain problems. It serves as a standard to compare learner’s knowledge to.
- The student diagnosis module collects information about learner’s knowledge and misunderstandings, creating the student model.
- The tutoring module holds teaching strategies and instructions to implement tutoring process. The primary tasks of this module are controlling selection, sequencing and presentation of learning material that is most suitable for the learner, determining the type and contents of feedback and help, and answering learners’ questions. Strategies contained by the module must be adapted to the needs of each individual learner without any help of humans. The goal of the module is to reduce the gap between learner’s knowledge and expert’s knowledge as far as possible or in ideal case eliminate the gap completely.
- The communication module is the only module interacting with the learner. It has to manage the user interface of the system. It perceives all learners’ actions, receives all requests from learners and forwards them to other modules. It is responsible for presentation of all kinds of information (curriculum of the course, materials, problems, feedback, etc.) to the learner, too.
Despite the acceptance of the modular architecture its main drawback is insufficient modularity to build complex adaptive ITSs because the modules have many tasks. Distributed computing technologies, namely, services and agents are used to split the higher level modules into lower level components to increase modularity of ITSs. Moreover, the functionality of ITSs includes many functions and corresponding pieces of code that differ in systems for various courses and even may be needed to change for the same course. Usage of distributed technologies also enables implementation of open ITSs allowing introduction of new functionality without changing existing code. The remainder of the chapter analyses use of the two types of distributed technologies in the architecture of ITSs.
- 1423 reads
