You are on page 1of 4

20i 0 2nd international Conference on Education Technology and Computer (ICETC)

The Application of AI Technology in Intelligent Tutoring System

DuLi Lin Zhuying* Yin Bing


College of Math and Computer Science College of Math and Computer Science College of Math and Computer Science
Guizhou Normal University Guizhou Normal University Guizhou Normal University
Guiyang, China Guiyang, China Guiyang, China
duli311@163.com zhuying_Iin@126.com andy-yb@163.com

Abstract- Intelligent Tutoring System (ITS) is the result of the replace the human teachers and to achieve the best teaching
application of artificial intelligence in education. This paper by letting the computer system be the guide and helper of
briefly introduced the concept of ITS; detailed its structure of learners, that to give your computer intelligence.
modules and features; and focused on the application of AI Researchers have different viewpoints on the structure of
technology; several trends of ITS development were analyzed ITS. More domestic scholars agree on the four modules
in the end.
theory namely, that a typical ITS system has four main
components: domain knowledge base, student model, tutor
Keywords-Artificial Intelligence (AI); Intelligent Tutoring
model and intelligent interface. (Shown as Figure 1)
System (ITS); Trends
...................

I. INTRODUCTION

Intelligent Tutoring System (ITS) is an important Student


Interface
research area in which computer science and technology are
used in education. It plays an important role in that with
Artificial Intelligence (AI) technology, in the absence of
guidance from the human tutors to help learners acquire
knowledge and skills. Since 1970s, many countries and
Teacher
regions have attached the great importance to intelligent Interface
tutoring systems research, development and application, and
had made great achievements. However, with in-depth
research, many problems began to surface. For example, in Domain
the traditional ITS, the student is the passive recipients of Knowledge Intelligent
teaching content. it lacks an interaction between systems and Interface
the students. In addition, the traditional ITS model of the
learners is not very successful, either too complicated to be Figure I ITS structure

handled, or too simple that it lacks practical value. Thus, in


the late of 1990s, ITS researchers began to reconsider the A. Domain Knowledge Base
application of artificial intelligence in ITS. Many new
technologies have been applied to ITS, including multimedia, As an important component of ITS, domain knowledge
Internet-based distributed education, hypermedia technology. base is to address the question of "what to teach" by
However, the introduction of new technology does not integrating multimedia database with artificial intelligence
reduce the importance of artificial intelligence in teaching, techniques. Its functions are as follow: as the source of all
on the contrary, these new technologies is of practical the knowledge of system, to be frequently called by other
significance only when combined with artificial intelligence. modules to fulfill user behavior response in real-time, by
This paper focuses on the application of artificial intelligence using the knowledge base to generate the appropriate issues,
technology in ITS by describing the structure and function of tasks, and interpretation; through synchronization questions
intelligent tutoring system, and discusses several and the comparison between expected student behavior and
development trends of ITS. actual student behavior to assess their mastery of knowledge
and learning status, learning style preferences, and so on.
II. THE MODEL OF BASIC STRUCTURE OF ITS Domain knowledge base can also lead the sharing of
teaching resources to carry out updating and refmement of
ITS is always the core content of artificial intelligence in
knowledge, in order to reduce a large number of
education (AIED), this system has the flexibility to represent
modification works of script design and maintenance. In fact,
the teaching materials according to the features of cognitive
knowledge representation is the primary problem which all
characteristics of students and current level of knowledge
the AI techniques and intelligent systems must tackle first.
and the ability to response to different needs of students,
realizing systematical intelligence by providing the teaching B. Student Model
strategies of "teaching according the information of Student model is the core component of ITS related to
students". Thus the ultimate goal of ITS research is to the issue of "who is learning". Its functions are as follow:
*Corresponding author

978-1-4244-6370-11$26.00 © 2010 IEEE VI-490


20i 0 2nd international Conference on Education Technology and Computer (ICETC)

According to the student model and domain knowledge base D. Intelligent Interface
to modeling the current student; to provide information for This is the modules that system communicates with users,
the decision-making of teaching modules; to maintain the human-machine interface is divided into teacher and student
established student model. Therefore, the student model is to interface, its function is that experts and users input
implement the personalization and adaptive teaching. With information were translated into the form of content that
the research of student model, a variety of student model system can accept, while experts and users input information
construction methods has been advanced. There are three were converted to the form of easy to understand for human.
common student models: Most of the teaching system using the graphical user
1) Overlay Model interface (GUI) or natural language communication interface,
This model is the knowledge of students as a subset of or both. Teaching system in the design of the user interface
expert knowledge. In view of this model does not take into should pay attention to the graphic simulation fidelity,
account the students may have a defect beyond the scope of user-friendliness and so on.
domain knowledge to produce the following correction
model. III. THE ApPLICATION OF AI TECHNOLOGY IN ITS
2) Deviation Model
The model can track the differences between students'
We have already discussed the main structure and
knowledge and domain knowledge, and register the students'
function of ITS, and now we look at several key technologies
knowledge blank and misconception. (Shown as Figure 2)
in which the application of AI.
Students' know ledge Domain Knowledge
A. Knowledge Representation provide strong support for
domain knowledge base
Knowledge organization was good or bad, which is the
success of ITS plays a decisive role. Good knowledge
representation can not only reasonable and effective manner
with the field of knowledge organization, but also provide
information for the system queries, can also help students
develop a clear concept of relationship, to establish a good
Figure2 Deviation Model structure of knowledge and improve problem-solving ability.
Knowledge blank is the knowledge in domain knowledge The knowledge in ITS can be divided into domain
base but the student does not know, and misconception is the knowledge (referring to the factual knowledge) and
error knowledge not in the domain knowledge base but the experiential knowledge (referring to the process of
students have. Another misconception is usually placed in a knowledge). Domain knowledge is mainly the course content
library, called the error database. to teach, experiential knowledge is teaching experience
3) Bayesian Networks which were accumulated by teachers in the teaching process.
Bayesian network based on students interact with the For these two types of knowledge, the former adopting a
intelligent tutoring system to speculate the statistics of cognitive hypertext knowledge representation, specifically
students knowledge. In the network each node has a the knowledge representation is composed of a data structure,
probability value to indicate the possibility that the students the knowledge technology is used to determine the
understand the corresponding knowledge point. knowledge elements, organizational computer access and
inference of knowledge of data structural form, and to design
e. Tutor Model based on the cognitive theory , namely hypertext theory; the
The module's main function is to how to organize the latter production rule-based knowledge representation
teaching content, that is, solve the problem of "how to teach". method, it can also defme sexual knowledge and process
Main functions: to provide students with the learning knowledge, and these two different characteristics of
environment, to guide students in learning activities, to knowledge can expressed in a unified form. In addition, there
explain the phenomena, processes and reasons, to provide are commonly used means of knowledge representation as
help and learning materials for students, monitoring and semantic network and logical notation. The former applies to
evaluating Students' learning activities. the defmition and expression of sexual knowledge, can easily
This module arrive intelligent teaching methods and describe the various physical entities of the semantic
teaching strategies according to the students' learning model relationships; the latter is very suitable for language teaching
and through the intelligent system of search and reasoning, system of knowledge representation, the logic of the
An ITS should encompass a variety of teaching strategies, mechanism can be used to operate all its content.
the system according to the needs of schools, choose a
B. Intelligent Agent technology applies to student model
different teaching strategies generated showing the relevant
and tutors model
parts of the teaching content and how to send the
decision-making. Intelligent Agent technology is a new and advanced
technology in the artificial intelligence field of computer
science. It is the entity which is able to dynamically

VI-491
20i 0 2nd international Conference on Education Technology and Computer (ICETC)

changing independent continuously, can operate in other nature of this transfonnation due to learning strategies the
agent of the environment, and can interact with the system used. Some basic learning strategies are memory
environment. learning, teaching learning, deductive learning, inductive
Students can use the agent as guidance to look for learning, analogical learning, etc. Some technologies in the
effective knowledge, since it has a learning function that can field of machine learning techniques can be applied to
fmd and collect the needed infonnation initiatively and intelligent teaching include: neural networks, learning based
efficiently from the Internet for users. In addition, in on statistical, inductive logic programming, data mining and
"intelligent navigation" browsing, you can also use so on.
"intelligent agent" technology to replace teachers to guide
E. Natural Language Understanding to facilitate
learning and search learning. Therefore, it helps to solve
man-machine interactive function
some problems as using a single keyword matching queries,
the search engines due to the emergence of a large number of In many ITS Intelligent interface module, the use of
irrelevant infonnation, infonnation retrieval accuracy not character recognition, speech recognition, natural language
high and so on. understanding technologies to allow students to use natural
Multi-agent technology is a revolutionary technology for language and "smart teachers" for man-machine dialogue.
ITS. In fact, teaching is a very complex process. It can not be Natural language understanding is that the students and the
a result of an isolated entity, but only able to rely on the interface part of the ITS. Can the students easily in natural
complementary functions of the interaction of mUltiple language questions to the intelligent tutoring systems, ITS
intelligent agent cooperation to complete. Therefore, ITS can give out the grammatical structure of natural language
based on distributed multi-agent has been a great application interpretation is ITS important issues to be resolved, in a
in the education of network. sense, natural language interface is the success of ITS
whether or not one of the keys.
C. The teaching strategy selection under the inference
In the intelligent interface module, the research results of
engine a variety of natural language processing technology can be
The main module of ITS is the tutor model, the role of combined to improve the system's intelligence level. For
the tutor model is mainly achieved by intelligent inference example, through natural language man-machine interface
engine , and the inference engine can be based on students more convenient to carry out human-computer interaction;
learning model provides learners in the past records and can use semantic web technology to achieve the level of the
current learning state to carry out reasoning, to select or relationship between the knowledge points and complexity
access to appropriate teaching strategies. The structure is of the semantics of links; can be intelligent fuzzy query
shown in Figure 3. technology knowledge base in search of knowledge and use;
can make use of machine translation technology, knowledge
Super rules of cross-language learning.

Choose Fixed
IV. DEVELOPMENT TREND
Match
With a variety of techniques and artificial intelligence
Rule storage
applications development in many fields the intelligent
tutoring systems are also constantly improved. Which the
Figure3 Inference Engine very worthy of attention are:
• To create ITS ontology to organize and promote the
The tutor model processing infonnation is inevitably concept of intelligent tutoring system
affected by some uncertainty factors as fuzziness, co-construction.
randomness, non-reliability and so on, so uncertainty • To develop intelligent tutoring system architecture
reasoning methods are adopted generally. Commonly
classification and mapping applications.
uncertainty reasoning methods are subjective bays methods, • Increase teaching agent and virtual humans' research.
non-numerical semantic description method, the possibility
• The promotion of team training research and
theoretical methods based on fuzzy set theory, belief network
prototype development.
approach and relational methods.
• The continued development of intelligent tutoring
D. Machine learning provides an automatic knowledge system and authoring system.
acquisition Apart from the above trend away from, online
The tutor models in any ITS will inevitably use the collaborative learning, serious games and interactive
machine learning techniques. For example, ITS want to narrative learning environments, adaptive learning support
summarize students' problem-solving style from a large systems are the intelligent tutoring system for opening up
number of answers, you must use machine learning new territories in recent years.
techniques. The purpose of machine learning is that the A. Support Online Collaborative Learning
infonnation in databases and infonnation systems will be
automatically extracted and converted into knowledge, and In recent years, with the development of E-Learning,
automatically added to the domain knowledge base. The support online collaborative learning become to a hot spot of
ITS research. Collaborative learning is the way that some

VI-492
2010 2nd International Conference on Education Technology and Computer (ICETC)

students to help each other to jointly solve the problems with adaptability. It is this ability of dynamic strain to the
no teacher intervention. The part of supporting online technology of artificial intelligence from the other separate
collaboration learning, ITS research is focused on from nature. ITS as a new teaching model, there are many
intelligence support to the team training. The research in this issues to be explored and resolved, should step up research
area in the early study of intelligent tutoring system is very efforts to learn the advanced experience at home and abroad,
limited, more in computer-supported collaborative learning from the actual situation, actively practice, dare to explore,
(CSCL) studies. On-line collaborative learning is usually so ITS can really play to their teaching function better serve
present with low participation rates and low exchange rates. the teaching service.
Therefore, to enhance visualization of the team, in support of
equality and mutual benefit teaching and team members are ACKNOWLEDGMENT

all intelligent matching, etc. need to be further solved. This work is supported by The Design of Intelligent
Educational Game in Network Course Learning Platform,
B. Serious Games and Narrative Learning Environment
under the project of Graduate Innovation Foundation of
Serious games (SG) are usually a kind of look and feel of Guizhou Normal University, and The Research of The
the game for the reality of the event or process simulation, Realistic Graphics Technology in Virtual Lab, under the
whose main purpose is to train or educate users. The project of Natural Science Foundation of Technology Office,
narrative learning environment (NLE) is a special kind of Guizhou Province.
narrative as the main means of communication with the
intelligent learning environment. SG or NLE with the REFERENCES:
advantage that a common learning environment does not [1] Chen Tianyun, Zhang Jianping. Intelligent Tutoring System (ITS) of
have: In order to provide learners with a real-world the status quo and its development in China [J]. China Audio-Visual
simulation and emulation, at the same time be able to Education, 2007, (241).

mobilize a large extent on the motivation of learners. ITS [2] Jiao Jialin, Xu Liangxian, Dai Kechang. Artificial Intelligence in the
research in this area is to give full play to SG or NLE the Intelligent Tutoring System [J). Computer Simulation, 2003, (8).

advantages of integration into a wealth of domain know ledge, [3 ] Nie Yonghong, Deng Guohe. ICAI exploration [J]. Journal of
Guangxi University of Technology, 2000, (12).
the learning process to enhance the user guide, so that they
[4] Shute, V. & Psotka, J. Intelligent Tutoring System: Past, Present,and
do not deviate from the direction of learning.
Future[A). DJonassen. Handbook of Research for Educational
C. Adaptive Learning Support System Communications and Technology[C]. New York: Macmillan.
[5] Lane, H. Chad.lntelligent Tutoring Systems: Prospects for Guided
With the popularity of constructivist learning theory, Practice and Efficient Learning [R).Whitepaper for the Army's
adaptive learning support gradually become a new research Science of Learning Workshop, Hampton, VA., 2006.
trend of ITS. Adaptive learning support system (ALSS) is [6] Murray, T., Blessing, S., & Ainsworth, S. (Eds.) Authoring Tools for
the intelligent tutoring systems and adaptive hypermedia Advanced Technology Learning Environments: Toward Costeffective
systems in a networked environment to develop integration. Adaptive, Interactive and Intelligent Educational Software
[M).Norwell, MA: Kluwer Academic Publishers, 2003.
Adaptive learning support system of pre-diagnostic tests and
learners in the learning process to update the model [7] Shute, V. J. & Psotka, J. Intelligent tutoring systems: Past, present,
and future[A]. D. H. Jonassen (Ed.), Handbook of Research for
information, the purpose is to grasp the learner's experience
Educational Communications and Technology[C). New York:
in real time based on cognitive structures, cognitive style, Macmillan. 1996: 570-600.
interests, hobbies, etc. of information, whereby in the current [8] Bloom, B.S. The 2 Sigma Problem: The Search for Methods of Group
state to provide learners to learn new information the most Instruction as Effective as One- to- One Tutoring [1]. Educational
suitable learning content, learning strategy, learning support Researcher, 1984, (13): 3-16.
and the best knowledge of the foundation. As the adaptive
learning support system for integration of an intelligent
"teaching" and the adaptation of the "learning" advantages,
but also help support the network environment adaptive
learning.

v. CONCLUSIONS

In the discussion that ITS is not difficult to be found, in


the ITS and its development prospects are involved in
artificial intelligence technology, artificial intelligence
technology will play a tremendous role in promoting
development. Should be focused on artificial intelligence is a
scientific tool for modeling the real world, this feature, make
good use of this tool to research, development and
manipulation able to demonstrate "teaching" and "study"
two essential features of the formal model, and that these
model is able to perceive the fact that the new learning and
reasoning, and ultimately has a human-like dynamic

Vl-493

You might also like