=Paper=
{{Paper
|id=None
|storemode=property
|title=Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems
|pdfUrl=https://ceur-ws.org/Vol-924/paper17.pdf
|volume=Vol-924
|dblpUrl=https://dblp.org/rec/conf/balt/Dabolins12
}}
==Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems==
https://ceur-ws.org/Vol-924/paper17.pdf
191
Trends of the Usage of Adaptive Learning
in Intelligent Tutoring Systems
Jānis DĀBOLIŅŠ 1
Faculty of Computer Science and Information Technology, Institute of Applied
Computer Science, Department of Systems Theory and Design, Riga Technical
University, Phone: +371 67089529
Abstract. In this paper general tendencies of adaptive learning are described,
which are becoming more and more common in the time when IT technologies and
use of internet as well as development of web systems is live topic in researches
about technology enchanted learning with the aid of intellectual learning systems.
Adaptive learning analysis, intellectual tutoring system description and general
ITS structure and general development principles are described based on the
bibliography. ITS modules, ITS collaboration with learner, necessity of feedback
and existing adaptive learning methods have been described.
Keywords. Intelligent tutoring systems, interactive learning environments,
adaptive learning, agent
Introduction
Computer usage in learning is connected with IT development in general; it began at
the end of 1950’s when machines, which are considered primitive nowadays, were
constructed for programmed training. Currently such technologies are used very widely,
for example, e-format information literature, virtual training systems and environments,
self-appraisal tests, technology enchanted learning as well as animated tutorials.
Improvement of IT technologies, expansion of internet and popularization of web
technologies have enabled technology enhanced learning introduction in adoption of
general matters and acquaintance of specialized problems. Thus researches on adoption
of learning contents into e-environment have become more necessary as a result of
such development [21, 9, 1]. Besides technology enhanced learning lets one to pick the
place and time where and when to study, which is great advantage compared to
traditional full-time education [16].
Identical educational tools and learning methods may not be effective or is less
effective for all students. It is possible to make the learning materials more flexible,
modify educational approach according to competence and temper of the student and
learning tasks, using Intelligent Tutoring Systems (ITS). It is possible to achieve better
learning results when fitted educational aids are used for individual needs of every
student [13]. Solutions are hunted in diversiform approaches, where ITS already takes
important place, as to a certain extent provides flexibility, adaptation, etc. Flexible
approach in learning is achieved by usage of several learning strategies, which are
1
janis.dabolins@rtu.lv
192 J. Dābolinš / Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems
implemented according to student’s progress in learning of the material. Such flexible
learning system for industrial purposes has already been described in 1998 [4], where
adaptive learning system is described with several learning strategies from several
agents.
Further researches [21, 13, 1] are focused on development of adaptive learning
system [11]. General adaptive learning tendencies are analysed in this paper and insight
onto further researches are given.
1. Analysis of Adaptive Learning
As volume of learning materials is great in e-environment, it is hard for consumers and
information seekers choose the right materials, thus it would be necessary to create
adaptive learning systems. Adaptive learning may be defined as ‘‘the process of
generating a unique learning experience for each learner based on the learner’s
personality, interests and performance in order to achieve goals such as learner
academic improvement, learner satisfaction, effective learning process and so forth”
[21].
Learning, methods and student's reaction to learning is not easily definable and
describable. Learning methods and successful learning results may not be defined as
particular thing, order of matters, sequence of events and thus guarantee successful
outcome (student is trained and knows everything that he/she needs). That is the reason
why adjustment of learning content to the student's competence is considered to be an
open system problem, where adaptation capacity of the student, collaboration with
learning environment as well as preferable result of system action plays great
importance (it is even hard to define the necessary outcomes, as it is not possible fully
evaluate persons’ gained knowledge and its system) [19]. In the same way it is hard to
define where system’s effect on person comes to an end.
When analysing person’s behaviour changes as a result of experience, comparing
student’s behaviour in time period 1 to behaviour in time period 2 (Figure 1) – if
behaviour is different in the same circumstances, we may consider that learning has
taken place. Such analysis of behaviour would give information on necessity of further
training for integrated learning agent.
In the development of learning systems it is necessary to take into account both
persons’ needs and requirements, as well as resources of information technologies. It is
also necessary to evaluate the use of didactic materials – cover learning theory matters,
develop training content plan, training methods and organizational forms, all these
operations would be arranged and subordinated according to their possible
Figure 1. Reproduction of learning [5]
J. Dābolinš / Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems 193
formalization forms for depicting and use in ITS. It is required to include student’s
reaction to the learning material and analyse the chances of learning outcome
stimulation.
In natural environment teacher may pick up each student’s individual abilities,
character, reaction to learnable material and depending on these parameters, form more
or less individualized task set for each individual. It is necessary to integrate analyse
mechanisms and reactions to imitate or overcome natural environment achievements.
On the other hand if the teacher is located in front of large student group, he/she is not
able to catch each student’s abilities and necessities, thus e-learning advantage is
silhouetted in this case. In e-learning system it is easier to analyse each individual
needs [1]. When adjusting training to individual needs, following aspects should be
included: individual intellectual abilities, perceptive type, preliminary knowledge,
desire for learning, motivation for achievements, self-suggestion [1]. Currently
researches are based on development of adaptive systems according to two approaches:
1. adaptivity – system structures didactic material combination according to knowledge
about the student; 2. adaptability – system adjusts learning content responding to
change in student’s knowledge during the use of system [21]. Studies are carried out on
development of adaptive learning materials, based on these two approaches.
2. Intelligent Tutoring Systems
Technology enchanted learning is learning where self-motivation, communication,
efficiency and technologies are used [16]. ITS is system and technologies where
adaptive learning technologies are used, which help individualize and personalize
learning process according to individual character and needs [8], analyse knowledge of
the theme, student’s mood and emotions (human decisions are based not only on
analysis of various possibilities and resulting signs, but also on emotions) [18, 2], as
well as learning style, typically ITS is constructed as multi-agent systems [3].
There are three main approaches on ITS development [13]:
1. The sequence of curriculum is made so that the student may easily adapt
himself/herself, besides material is demonstrated to the student just when
he/she needs it.
2. ITS gives detailed feedback to the learner on the imperfect or false solution,
helping to learn from ones mistakes.
3. Problem solving methods - little help is provided to the learner, so the right
solution is achieved.
Figure 2. General ITS structure
194 J. Dābolinš / Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems
General ITS structure is given in Figure 2. It is possible to create constructive
multi-agent system for adaptive learning, when describing and formalizing information
on person’s knowledge accumulation and interpretation types. ITS includes both
student/learner module and expert module, both of those mutually interacts with
tutoring module, which is available for users in user interface. Operation of agents and
adaptation to learner takes place in learning/tutoring module, where system receives
information about user’s activities, acquired knowledge and skills [18, 16, 14].
The necessary information for this system may be supplemented in the expert
module, but the information about student is stored in the learner module. Both of these
modules interactively serve as information storage for agents, thus operation of
modules is defined as follows:
• Learner module is accumulative type – it stores information about student’s
actions, progress and test results.
• Expert module contains information about learning content. It may be
supplemented and in collaboration with student module perform changes in
order to bring together learning module action with solving of students’
problems in possibly short period of time.
• Tutoring module cooperates with both previous mentioned modules for
gaining the material as well as determines the accuracy of student’s solutions,
based on information contained in the expert module. Module contains
algorithms as well, which performs help function for student during learning
process.
• Operation in e-learning system is done in the user interface.
Agents integrated in ITS should be able to express attention, adapt themselves to
abilities, learning speed, needs and necessities of the leaner [12]. These qualities are
attained by integration of fuzzy logic elements into the agent, using their response in
planning and expression of emotions in reaction to user’s activities.
To achieve high efficiency ITS collaboration with student, it is necessary to
integrate natural language processing (NLP) techniques, as a result agents integrated in
the system may analyse student's answers qualitatively [15]. If feedback between ITS
and learner is provided, better training results are achieved [10, 3, 17], if ITS instantly
corrects the mistakes made by learner or shows the right way how to avoid the mistake,
knowledge of the student becomes deeper and wider [20, 9]. The knowledge
assessment agent using the comparison algorithm based on graph patterns compares a
teacher’s and a learner’s concept maps and assigns score for submitted solution [7].
IKAS (Intelligent Knowledge Assessment System – developed at the Department of
Systems Theory and Design of Riga Technical University) basic concept is exercises
for knowledge assessment, where natural language processing is not necessary for
computerized assessment, as well as teaching staff involvement in evaluation process
of competence. Idea map exercises are different and thus are adjustable to different
knowledge assessment needs (more simple and complicated exercises). Besides variety
of exercises founded in idea maps and arranging possibility after level of difficulty,
allows to adjust them to adaptive knowledge assessment [3]. But the system
COMPASS (COncept MaP ASSessment and learning environment) is a discipline-
independent concept mapping learning environment, developed at the Educational &
Language Technology Laboratory of the Department of Informatics &
Telecommunications at the University of Athens. The analysis of the map is based on
J. Dābolinš / Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems 195
the assessment of the propositions according to specific criteria, such as completeness,
accuracy, superfluity, missing out and non-recognisability, results into the
identification of specific error categories (e. g. incomplete relationship, incorrect
concept, superfluous relationship, missing proposition), and is discriminated in the
qualitative and quantitative analysis. The qualitative analysis is based on the qualitative
characterization of the errors and aims to contribute to the qualitative diagnosis of
student’s knowledge; that is student’s incomplete understanding/ beliefs and false
beliefs. The quantitative analysis aims to evaluate student’s knowledge level and is
based on the weights assigned to each error category as well as to each concept and
proposition that appear on expert map, reflecting their degree of importance. Pre-
defined weights for error categories are supported; the teacher has the possibility to
personalize the assessment process and configure the weights [6].
3. Conclusions
In former studies e-learning systems mostly have been formed as systems for
information supply, where lectures or distance learning materials are available in e-
format. Popularity of internet and web technologies have made e-learning development
possible, though available operating systems still are “teacher’s” systems where only
little attention is paid to learner’s needs and abilities [13]. In the latest researches target
has been set to individualized e-learning system, directed to the student. Technology
enhanced learning have adaptation gaps, where one of the solutions is ITS – where
realization of adaptation is made, so that ITS is brought nearer to the person’s ability to
learn. Currently feedback role is developed the most in ITS, IKAS and COMPASS.
When analysing existing ITS we may conclude that their development is
progressing, but still it is not at the level where ITS fully realizes adaptive training –
technologies which let course to comply with students’ knowledge level and
preferences automatically. Adjustment of learning content and didactic materials to
student’s needs is object of many researches, but trends differs – several researches
tries to develop standardized learning content, which could be modified according to
progress of the student, others in their turn are directed to personalization of learning
content, adaptation directly to student's goals and achievable results. All of these
approaches have their drawbacks – first, most of the results and elaboration of these
researches are not compatible with existing learning standards, second – ITS that
adjusts to the interests of learner may omit important parts of the learning content.
Irrespective ITS drawbacks, elaborating and research is very current topic because of e-
learning privileges, thereby we may consider development in the direction of ITS to
overcome their imperfections and drawing nearer to real setting learning or even
gaining advantage over these.
Acknowledgement
This work has been supported by the European Social Fund within the project “Support
for the implementation of doctoral studies at Riga Technical University”.
196 J. Dābolinš / Trends of the Usage of Adaptive Learning in Intelligent Tutoring Systems
References
[1] Y. Akbulut and Ç. S. Çardak, Adaptive educational hypermedia accommodating learning styles: a
content analysis of publications from 2000 to 2011, Computers & Education 58(2) (2012), 835-842.
[2] E. Alepis and M. Virvou, Automatic generation of emotions in tutoring agents for affective e-learning in
medical education, Expert Systems with Applications 38(8) (2011), 9840-9847.
[3] A. Anohina-Naumeca and J. Grundspeņķis, Evaluating students' concept maps in the concept map based
intelligent knowledge assessment system. In: Advances in Databases and Information Systems,
Associated Workshops and Doctoral Consortium of the 13th East European Conference, 7-10 September
2009, Riga, Latvia. Lecture Notes in Computer Science 5968, Springer, Berlin, 2010, 8-15.
[4] C. Frasson and E. Aimeur, Designing a multi-strategic intelligent tutoring system for training in industry,
Computers in Industry 37(2) (1998), 153–167.
[5] N. L. Gage and D. C. Berliner (N. L. Geidžs and D. C. Berliners), Pedagoģiskā psiholoģija/Educational
Psychology, Zvaigzne ABC, Riga, 1999.
[6] E. Gouli et al., Exploiting COMPASS as a tool for teaching and learning, Proceedings of the 3rd
International Conference on Concept Mapping, Tallinn, Estonia & Helsinki, Finland, 2 (2008), 383-390.
[7] J. Grundspeņķis, Usage experience and student feedback driven extension of functionality of concept
map based intelligent knowledge assessment system, Communication & Cognition 43 (1-2) (2010), 13-
32.
[8] K. Günel and R. Aşlıyan, Extracting learning concepts from educational texts in intelligent tutoring
systems automatically, Expert Systems with Applications 37(7) (2010), 5017-5022.
[9] F. Gutierrez and J. Atkinson, Adaptive feedback selection for intelligent tutoring systems, Expert
Systems with Applications 38(5) (2011), 6146-6152.
[10] Y. He, S. Ch. Hui, T. T. Quan, Automatic summary assessment for intelligent tutoring systems,
Computers & Education 53(3) (2009), 890-899.
[11] B. G. Johnson, F. Phillips, and L. G. Chase, An intelligent tutoring system for the accounting cycle:
enhancing textbook homework with artificial intelligence, Journal of Accounting Education 27(1) (2009),
30-39.
[12] B. Kort, R. Reilly, and R. W. Picard, An affective model of interplay between emotions and learning:
reengineering educational pedagogy – building a learning companion. In: Proceedings of IEEE
International Conference on Advanced Learning Technology: Issues, Achievements and Challenges, 6-8
August 2001, Massachusetts, USA. IEEE Computer Society, 2001, 43-48.
[13] A. Latham, K. Crockett, D. McLean, and B. Edmonds, A conversational intelligent tutoring system to
automatically predict learning styles, Computers & Education 59(1) (2012), 95-109.
[14] P. J. Muñoz-Merino, M. F. Molina, M. Muñoz-Organero, and C. D. Kloos, An adaptive and innovative
question-driven competition-based intelligent tutoring system for learning, Expert Systems with
Applications 39(8) (2012), 6932-6948.
[15] V. L. Payne, O. Medvedeva, E. Legowski, M. Castine, E. Tseytlin, D. Jukic, and R. S. Crowley, Effect
of a limited-enforcement intelligent tutoring system in dermatopathology on student errors, goals and
solution paths, Artificial Intelligence in Medicine 47(3) (2009), 175-197.
[16] P. Phobun and J. Vicheanpanya, Adaptive intelligent tutoring systems for e-learning systems, Procedia -
Social and Behavioral Sciences 2(2) (2010), 4064-4069.
[17] I. Roll, V. Aleven, B. M. McLaren, and K. R. Koedinger, Improving students’ help-seeking skills using
metacognitive feedback in an intelligent tutoring system, Learning and Instruction 21(2) (2011), 267-
280.
[18] A. Sarrafzadeh, A. Samuel, D. Farhad, F. Chao, and B. Abbas, “How do you know that I don’t
understand?” A look at the future of intelligent tutoring systems, Computers in Human Behavior 24(4)
(2008), 1342-1363.
[19] S. Schiaffino, P. Garcia, and A. Amandi, eTeacher: Providing personalized assistance to e-learning
students, Computers & Education 51(4) (2008), 1744-1754.
[20] C. Woo Woo, M. W. Evens, R. Freedman, M. Glass, L. S. Shim, Y. Zhang, Y. Zhou, and J. Michael, An
intelligent tutoring system that generates a natural language dialogue using dynamic multi-level planning,
Artificial Intelligence in Medicine 38(1) (2006), 25-46.
[21] M. Yaghmaie and A. Bahreininejad, A context-aware adaptive learning system using agents, Expert
Systems with Application 38(4) (2011), 3280-3286.