Soutenance mémoire
Agent and Multi-Agent System
The use of Agent in Intelligent Tutoring System ( ITS )
With the social development, Intelligent Computer Assisted Instruction System, also called the Intelligent Tutoring System (ITS), has aroused much attention. Many kinds of ITS have been developed home and abroad. Network engineers from all other countries are conducting wide and further research on ITS and have indeed made great achievements.
Nevertheless, in China, the development of ITS lags far behind and cannot provide effective teaching. This leads to the low effectiveness and less of attraction of the network teaching system. Therefore, the task of top priority now is to improve the individualization and intelligentization of ITS.
In this thesis, the elaboration of agent and multi-agent system, the popular techniques in the field of Artificial Intelligence (AI) has been given priority. There is also a detailed introduction about the features and application of these techniques, hi 2.5, a conclusion that the use of agent techniques may improve the intelligentization of ITS has been reached.
This thesis tries to make full use of the multi-agent system to put forward a Distributed Multi Agent System (DMAS). All this is based on features of agents, whose cooperative relations can help solve the complicated problems in ITS. The application of the DMAS in the current network teaching can help greatly improve the ITS and promote its intelligentization. The system thus will be more humanized and individualized.
What is ITS?
Artificial intelligence in education comes of age in systems now called « intelligent tutors », a step beyond traditional computer-assisted instruction. Computer-assisted instruction evolves toward intelligent tutoring systems (ITS) by passing three tests of intelligence. First, the subject matter, or domain, must be « known » to the computer system well enough for this embedded expert to draw inferences or solve problems in the domain.
Second, the system must be able to deduce a learner’s approximation of that knowledge. Third, the tutorial strategy or pedagogy must be intelligent in that the « instructor in the box » can implement strategies to reduce the difference between expert and student performance.
ITS covers Artificial Intelligence, computer science, education, psychology, behavioral science, and many other fields. The study of ITS is to use computers more widely in education and realize the intelligentization of a computer system. In this way, the computer can replace teachers in some degree and help to optimize teaching and learning.
ITS is a comprehensive subject and is closely related to AI, computer science, cognitive science, noetic science, education, psychology and behavioral science. Its final aim is to make the computer shoulder the responsibility of human education, namely, to realize the intelligentization of a computer system and optimize teaching and learning by replacing teachers. The significance of the study of ITS is to reduce teachers’ amount of work and reach high standards of teaching. This requires that the computer system consist of domain knowledge, teaching knowledge, the capabilities of learner-computer interaction.
The major features of this system are as follows:
(1) It can automatically provide all kinds of exercises and rehearsal practices.
(2) It can adjust the learning content and rate of progress according to the abilities of different learners.
(3) It can solve problems on the basis of its comprehension of the teaching materials.
(4) It can understand and respond to natural language, thus form a free question-answering system and improve the interaction between learners and computer.
(5) It possesses the ability to explain the teaching materials.
(6) It can judge learners’ mistakes, analyze the reasons and correct it in time.
(7) It can appraise learners’ learning behavior.
It is difficult for the ITS to possess all the features mentioned above though it will be perfect to do so. Therefore, ITS need merely possess one or several of the features required.
Use in Practice
All this is a substantial amount of work, even if authoring tools have become available to ease the taskf . This means that building an ITS is an option only in situations in which they, in spite of their relatively high development costs, still reduce the overall costs through reducing the need for human instructors or sufficiently boosting overall productivity. Such situations occur when large groups need to be tutored simultaneously or many replicated tutoring efforts are needed. Cases in point are technical training situations such as training of military recruits and high school mathematics. One specific type of intelligent tutoring system, Cognitive Tutors, has been incorporated into mathematics curricula in a substantial number of United States high schools, producing improved student learning outcomes on final exams and standardized tests . Intelligent tutoring systems have been constructed to help students learn geography, circuits, medical diagnosis, computer programming, mathematics, physics, genetics, chemistry, etc.
In the following two chapters, a further discussion will be made about agent technology and multi-agent system
ITS Conference
The Intelligent Tutoring Systems conference was typically held every other year in Montréal (Canada) by Claude Frasson and Grilles Gauthier in 1988, 1992, 1996 and 2000; in San Antonio (US) by Carol Redfield and Valerie Shute in 1998; in Biarritz (France) and San Sebastian (Spain) by Guy Gouardères and Stefano Cerri in 2002; in Maceio (Brazil) by Rosa Maria Vicari and Fâbio Paraguaçu in 2004; in Jhongli (Taiwan) by Tak-Wai Chan in 2006. The conference was recently back in Montreal in 2008 (for its 20th anniversary) by Roger Nkambou and Susanne Lajoie. ITS’2010 will be held in Pittsburgh (US) by Jack Mostow and Vincent Aleven. The International Artificial Intelligence in Education (AIED) Societypublishes. The International Journal of Artificial Intelligence in Education (IJAIED) and produces the International Conference on Artificial Intelligence in Education every odd numbered year. The American Association of Artificial Intelligence (AAAI) (www.aaai.org) sometimes has symposia and papers related to intelligent tutoring systems. A number of books have been written on ITS including three published by Lawrence Erlbaum Associates.
IA in Distance teaching system
In artificial intelligence, an intelligent agent (IA) is an autonomous entity which observes and acts upon an environment (i.e. it is an agent) and directs its activity towards achieving goals (i.e. it is rational)ntelligent agents may also learn or use knowledge to achieve their goals.
Individualization of Distance teaching:
At present, individualization in distance learning assistant only focuses on time and space. The teaching methods and materials have not been changed virtually. Individual education can be regarded as one of virtues of the Web. However, if just change the classroom teaching into the form of network teaching, individual education can not be really carried out and the virtues of the traditional teaching method are lost. When agent technology is adopted, Intelligent Agent ( IA ) can choose different teaching methods and teaching resources for each learner. This selection is based on learners’ various learning levels, teaching materials, difficulties in learning and learning motivations. Each learner will experience a different learning process. The agents used in the system will become « private tutors » for each learner.
Human-computer interaction and teaching method:
With the introduction of agent technology, the human-computer interaction will undergo a radical transformation. The essential feature of the network teaching based on multi-agent is its great « kindness » towards learners. During the teaching process, learners will study in a completely different environment, i.e., the computer can « hear » the learners’ voice and accordingly adjust the whole teaching process.
Self-adaptive evolution of agent:
When self-adaptive evolution of agent is introduced into education, the potentiality of the system is optimized. The agents provide different learning patterns to each learner. But they will continue to change in accordance with the human-computer interaction. Deep acquaintance with the learner will make the system more adaptable to learners’ requirements. In some degree, the quantity and quality of the teaching resources mean the adaptability of the system to learners.
Efficient use of teaching resources:
At present, teaching resources available on the Internet are in abundance. However, they are almost in a mess and can only be regarded as some kind of information or data.
What learners want is knowledge. It is a great, complex task to transform all these information and data into useful knowledge topics. In the current situation, many learners are « drowned » when they try to utilize the resources on the Internet. They just cannot use the resources efficiently. In ITS based on multi-agent, the agents will help transform what is need by learners. Accordingly, all the information and data available on the Internet will be fully used by learners.
Cooperation and intelligentization:
In ITS based on multi-agent, questions in the discuss board will be summed up and special topics will be further discussed. Learners can understand those questions more thoroughly. This shows the virtues of cooperation in ITS.
Networking teaching helps cultivate learners’ abilities and develop quality education:
Learners may bring forward their own questions. With the help of ITS, they can solve their problems step by step according to what have been discussed. They can even further their research. Instead of a passive teaching model, ITS based on multi-agent provides exhaustive resources and an open environment. Learners themselves can decide which part is more important in their study and which learning pattern they prefer. In the learning process, all questions will be solved gradually and learners’ abilities in solving problems will be improved at the same time.
All in all, ITS base on multi-agent has a promising prospect in network teaching.
This thesis mainly focuses on the application of agent technology in ITS. More details are provided in the following part.
Agent technology is a hot topic in AI at present. People have been trying to use the agent technology to unify and further develop AI. Some even try to use it to unify and develop software. This shows that agent technology has a wide and promising application prospect. Thus in the application of ITS, we can utilize the multi-agent system. This is mainly because the independent and cooperative relations between agents feature largely the multi-agent system. And these features will help solve the complicated problems in ITS and support the realization of the complex functions of the teaching system.
In the following two chapters, a further discussion will be made about agent technology and multi-agent system.
|
Table des matières
CHAPTER 1 :Introduction
1.1 Background and Purpose of Research
1.2 The use of Agent in Intelligent Tutoring System ( ITS )
1.3 Thesis Organization
CHAPTER 2 :Intelligent Tutoring System
2.1 What is ITS?
2.2 The history of ITS
2.3 Use in Practice
2.4 ITS Conference
2.5 IA in Distance teaching system
CHAPTER 3 :Agent and Multi-Agent System
3.1 Background of Agent
3.2 Basic concepts concerning agents
3.2.1 Definition of agents
3.2.2 Features of an agent
3.2.3 The architecture of an agent
3.3 The Multi-Agent System
3.3.1 The concepts of Multi-Agent System
3.3.2 The architecture of Multi-Agent System
3.3.3 The application of MAS
3.3.4 Types of MAS
CHAPTER 4 :Multi-Agent System
4.1 Knowledge presentation and reasoning
4.2 Agent Communication Language (ACL)
4.2.1 Features of Agent communication language
4.2.2 Knowledge Query Manipulation Language (KQML)
4.3 The agent-based communication mechanism
4.4 Multi-agent coordination models
4.4.1 Definition of multi-agent coordination models
4.4.2 Classification of multi-agent coordination models
4.5 Multi-agent negotiation models
4.6 Task decomposing and scheduling
CHAPTER 5 :Multi-Agent System (MAS) model
5.1 Model of MAS used in application system
5.2 The architecture based on the model
5.2.1 Architectural structure
5.2.2 User interface design
5.2.3 User agent design
5.2.4 Application agent design
5.2.5 Development of the Control Agent
CHAPTER 6 :The application example of Multi-Agent System — Intelligent Tutoring System on Network
6.1 Overview
6.2 CAI Modeling
6.2.1 Components of the System
6.2.2 System structure
6.2.3 Tutoring Agent
6.2.4 The System Database
6.3 Network tutoring process instance
6.3.1 User agent-student agent
6.3.2 User Agent Teacher Agent
CHAPTER 7 :Conclusion
Télécharger le rapport complet
