Towards an Adaptive Intelligent Assessment Framework for

Collaborative Learning

Asma Hadyaoui and Lilia Cheniti-Belcadhi

Sousse University, ISITC, PRINCE Research Laboratory, Hammam Sousse, Tunisia

Keywords: Computer-Supported Collaborative Learning, Assessment, Adaptation, Assessment Model, Group Learner

Model, IMS/QTI, Ontology.

Abstract: Assessing in an online collaborative learning environment is a complex task due to the variety of elements

and factors that intervene in how a group of learners collaborates to achieve an assessment task. This paper

aims to improve both learners’ and group performance at a given activity or a set of activities by adapting the

assessment process to the learner level. To that end, we propose a general framework to illustrate our adaptive

approach for assessment in an online collaborative learning environment. To do so, we take the concept of

adaptation, generally based on three models: the learner model, the domain model, and the adaptive model,

as a point of departure and extend it by designing two other new models that are an assessment model and a

group learner model. To present our assessment model, we are based on IMS/QTI standard and ontology for

the formalization of the question. We aim to combine collaborative learning, assessment, and adaptation to

provide an adaptive assessment, an adaptive group composition, and an adaptive collaboration.

1 INTRODUCTION

Nowadays, learners use technology anyplace,

anytime, as a consequence, they require adequate

learning types that are challenging and engaging

(Mariel Miller, Allyson Hadwin, 2015). The use of

information and communication technologies in the

education sector for learning and assessment in

various forms has been the object of intense research

for several years (Gembarski, Paul., 2020). Within the

new era for assessment, learners are provided with

timely and quality feedback to scaffold their learning

process and to maintain their progress and success

(Susan Finger, Dana Gelman, Anne Fay, Michael

Szczerban, 2006). Learners play major roles in the

assessment process as they participate in alternative

forms of assessment based on their behavior and

performance. However, traditional instructor-

centered examination remains the primary means for

assessing learner performance, and collaborative

learning is undervalued and marginalized. In a large

part, this is because the assessment of collaboration

requires new approaches and methodologies. In this

paper, we focus on how to make the assessment

process intelligent and personalized. Doing so, we

propose an intelligent assessment framework in an

online collaborative learning environment. It is

paramount to provide intelligent real-time feedback

while performing collaborative tasks and to analyze

the behavior characteristics of online learners to

intelligently adapt online assessment strategies and

enhance the quality of learning. In previous research

work, we have already focused on personalized

feedback generation in online learning Environments

(Belcadhi, 2016). The underlying problem is: How to

make our assessment strategy intelligent and

personalized to learner and group profile,

performance level, and preferences, in an online

collaborative learning environment?

The first problem here is the collaboration itself,

how a system can effectively support collaboration

patterns, and how it can be aware of the current

collaboration status? To overcome this issue, our

framework will be based on Computer-Supported

Collaborative Learning (CSCL). It provides the

possibility of learning through collaborative

interaction, and the social construction of knowledge

through the utilization of information technology

(Allaymoun M. H., 2021).

For the personalization challenge, adaptation is

often proposed as a way of overcoming it. In our

context, the assessment process has to meet the

groups and learners’ levels and preferences. Adaptive

Hadyaoui, A. and Cheniti-Belcadhi, L.

Towards an Adaptive Intelligent Assessment Framework for Collaborative Learning.

DOI: 10.5220/0011124400003182

In Proceedings of the 14th International Conference on Computer Supported Education (CSEDU 2022) - Volume 1, pages 601-608

ISBN: 978-989-758-562-3; ISSN: 2184-5026

601

tutoring systems can consider learner characteristics

such as knowledge, affective state, and learning style

as a basis for providing adaptation. In our case, we are

more interested in the profile, level, and activities

preferences for both individual and groups of

learners.

This paper contributes to the assessment, CSCL,

and intelligent tutoring system. We propose an

approach for the adaptation and personalization of the

assessment task and instruction offered to learners

and groups in an e-learning environment. Changing

assessment activities based on individual and group’s

performance on the previous task: the difficulty of

activities will increase as a learner does them

accurately, while if the student struggles the tasks get

easier. To do that, we are based on the classical

architecture for the adaptive educational system, we

extended it by adding both an assessment model and

a group learner model. Our main contribution for

these areas is a conceptual framework, which

proposes an original approach to the integration of the

intelligent adaptive aspect inside the assessment

process in an online collaborative learning

environment. To our knowledge, there is no system

whose development is carried out in this perspective,

and this would therefore constitute the originality of

the approach conducted in this paper.

The rest of this paper is organized as follows.

First, we mention some concepts and works related to

our research. Then, we present our approach for the

proposed adaptive collaborative assessment system

including a general conceptual scheme for the

adaptive final intelligent system and a meta-model for

the assessment platform. Finally, conclusions and

future research work are outlined.

2 LITERATURE REVIEW

In this paper, we propose to tackle assessment in an

Online Collaborative Learning Environment. In such

an environment, students learn in groups via

interactions with each other by asking questions,

justifying their opinions, explaining their reasoning,

and presenting their knowledge (Soller, 2001).

2.1 Collaborative Learning

Collaborative learning is often defined as “a situation

in which two or more people learn or attempt to learn

something together, and collaboration involves the

mutual engagement of participants in a coordinated

effort to solve problems together” (Dillenbourg,

1999). Several researchers have pointed out the

importance of a collaborative environment and how

significantly effective it is in terms of learning gain

(Hayashi, 2014). it's important to tell apart

collaboration and cooperation within the method

participants perform the actions to a shared objective:

“collaboration means the work is accomplished by all

the participants together; whereas cooperation means

that participants act towards a shared goal, but each

of them performs specific and independent actions to

achieve part of the overall goal” (Jeremy Roschelle

and Stephanie D Teasley, 1995). Rapid developments

in computer-mediated communication in the late

1980s led to a new discipline in the 1990s now

referred to as CSCL (Lipponen, 2002). It is also one

of the most important computer-supported learning

fields that improve learning and employ collaborative

work to enable learners to discuss their ideas and

present their views, allowing the exchange of ideas

and information (Lipponen, 2002).

2.2 Computer-Supported Collaborative

Learning (CSCL)

CSCL is a fundamental paradigm that uses

information technology tools that help to learn

processes (Allaymoun H. , 2014). It is one of the most

promising innovations to enhance teaching and

learning using ICT tools. It includes a range of

situations in which interactions take place among

students using computer networks to improve the

learning environment. The primary aim of CSCL is to

provide an environment that supports collaboration

between students to improve their learning processes

(Karel Kreijns, 2003) and facilitate collective

learning (Pea., 1994) or group cognition (Stahl,

2006).

Collaboration is a complex activity that involves

both individual and group effort. To encourage

collaboration, each aspect should be assessed

(David.W. Johnson and Roger.T. Johnson, 1992).

The way to ensure individual accountability, in which

students are held responsible for their learning, and

positive interdependence, in which students reach

their goals if and only if the other students in the

learning group also reach theirs, according to Johnson

and Johnson, is to assess both individual and group

learning (David.W. Johnson and Roger.T. Johnson,

1992). To show the benefits obtained from using

CSCL, especially in tutoring systems, Kumar and

Rose, in 2011, built intelligent interactive tutoring

systems CycleTalk and WrenchTalk that support

collaborative learning environments in the

engineering domain (Conati, 2009). According to

Rohit Kumar and Carolyn. P. Rosé, students who

EKM 2022 - 5th Special Session on Educational Knowledge Management

602

worked in pairs learned better than students who

worked individually (Rohit Kumar, Carolyn P. Rosé,

2011) (Rohit Kumar, Carolyn. P. Rosé, Yi-Chia.

Wang, Mahesh Joshi, and Allen Robinson, 2007).

Another tutoring system that supports collaborative

learning is described in (Jennifer K. Olsen, Daniel M.

Belenky, Vincent Aleven, Nikol Rummel, 2014) for

teaching mathematical fractions.

2.3 Adaptation

Adaptation in learning systems can be defined as the

ability of a system to adjust instructions according to

the learners' abilities and/or preferences. It considers

that identification of learning style is a very crucial

tool to improve the individual learning of a learner,

especially in online learning (Bhawna Dhupia,

Abdalla Alameen, 2019). The objective is to act on

the identified characteristics of the learner and to

improve the effectiveness and efficiency of learning

(Steven Oxman and William Wong, 2014). The basis

of adaptive learning can be identified by three

elements shared by all adaptive systems (Peter

Brusilovsky, 2000). The learner model represents the

source of the adaptation, the domain model describes

the adaptation target, and the connection established

between the learner model and the domain model is

implemented by the adaptation model. The

introduction of other components in the system

architecture is also possible (Blake, Robert J., 2009).

Nevertheless, the three models mentioned above are

a necessary precondition for each adaptive system to

identify the individual characteristics of the learner

and to decide which, when, and how an adaptive

instruction will be delivered to a particular learner.

Adaptation provides adaptivity in terms of goals,

preferences, and knowledge of individual students

during interaction with the system. In our case, we

seek to filter and order assessment activities given to

learners. Rather than offering the same activity to

everyone, our approach aims to select the next

assessment activity to perform a certain learner based

on the previous performance. This allows adaptation

at each stage of the assessment process. We attempt

also to use knowledge about collaborating peers and

group interactions represented in the learner model

and group learner model to form a matching group for

different kinds of collaborative tasks. This is a new

approach expanding ideas from the classic adaptive

system, CSCL, and assessment to make an adaptive

collaborative framework for assessment in an online

collaborative learning environment.

3 ASSESSMENT FRAMEWORK

This section discusses the proposed framework to

implement our adaptive approach for assessment in

an online collaborative environment. The adaptive

assessment framework will be able to cater to the

needs of the heterogeneous type of users.

There are three basic models for an adaptive

learning environment namely, domain model, learner

model, and group model. To these models, we

propose to add the assessment model and the group

learner model. Therefore, the proposed adaptive

assessment framework revolves around five models.

The basic element of the proposed framework is the

assessment platform. This is primarily a source for all

types of data required for the system to carry forward.

It includes all the information relevant to the

assessment process with all the elements. The

assessment platform will provide assessment

indicators calculated while performing assessment

activities for all, the learner model, the group learner

model, the domain model, and the adaptive model.

Figure 1: The general scheme for the adaptive framework.

3.1 Assessment Framework

Description

To define the communication interface for the

adaptive assessment framework we propose a meta-

model using the Unified Modelling Language

(UML), a standard language for specifying,

visualizing, constructing, and documenting concepts

and artifacts. Specifically, the modeling elements

from the Class Diagrams of the language are used.

The meta-model for the Assessment platform is

shown in Figure 2. The different elements of the

meta-model are detailed next.

The concept of activity is the basic abstract

concept of the platform. An activity is a structured

object possibly containing an arbitrary number of

assessment tests. Activities are executed by an actor,

which can be a single person, a group of persons.

Activities also have Resources, representing elements

created or manipulated by the activity. Resources are

Towards an Adaptive Intelligent Assessment Framework for Collaborative Learning

603

used to perform an activity, learning outcomes, or

learning goals to be evaluated while performing the

activity. Indicators will be calculated based on

individuals’ and groups’ traces tracked while

performing assessment activities. An indicator is a

significant element, identified using a set of data, that

makes it possible to evaluate a situation, a process, a

product, etc. According to (Angelique, 2004) an

indicator is “a mathematical variable that has a list of

characteristics. It is a variable that takes values

represented by digital, alphanumerical, or graphical

forms. The value has a status: the value is calibrated

to other variables”. A lot of works is revealed

regarding indicators, typically respecting this

definition. For example, (Olga C. Santos, Antonio R.

Anaya, Elena Gaudioso, Jesus G. Boticario, 2003)

offers a tool that calculates from the interactions, the

degree of involvement of each learner during the

learning unit. It identifies participative learners,

useful learners, non-collaborative learners, learners

who take initiative, and communicative learners.

Figure 2: Meta-model for the assessment platform.

3.2 The Domain Model

The domain model includes a representation of the

knowledge and expert skills of a domain that can be

transmitted by didactic and pedagogical methods

(Nwana, H. S, 1990). Depending on the domain,

knowledge modeling will be performed. This

knowledge and skills are unpacked as knowledge

components (KC) in a way that facilitates their

representation as facts, principles, or rules according

to a hierarchy generally based on their complexity

(Tardif, 1999), and in a way that facilitates an

incremental acquisition in a learning process (John

Whiting and David. Bell , 1987). The domain model

provides our system with a baseline for inferring the

knowledge state of the learner or groups of learners.

3.3 The Learner Model

The learner model is defined by (Millán, 2007) as a

representation of information about an individual user

that is essential for an adaptive system to provide the

adaptation effect, i.e., to behave differently for

different users. The learner model is a representation

of the learner profile deduced from the assessment

activities. It is responsible for discovering the

individual learning behavior of the learner (Abdalla

Alameen, 2019). The learner profile can be conceived

at the epistemic level and the behavioral level

(Wenger, 1987). It aims to identify the individual

characteristics of each learner’s strengths,

preferences, and motivations (Hongchao Peng,

Shanshan Ma, Jonathan Michael Spector, 2019). At

the epistemic level, the data collected in the learning

environment is used to infer the learner's knowledge

status. This includes theoretical and declarative

knowledge, as well as procedural knowledge

(Wenger, 1987).

The updating of learner profile means the

updating of values associated with the Concept

Competence. It is based on the learner’s

performances on the pedagogical resources of type

test (exercise, problems, MCQ, question, etc). The

process of updating the learner’s profile is necessary

to keep track of the learner’s evolving competencies.

This updating affects the accuracy of the pedagogical

assessment activities proposed to the learner, which

in turn will help increase the learner’s performance.

3.4 The Adaptive Model

The learner model, the group learner model, the

domain model, and the assessment model are sources

for all types of data required for the adaptive model

to carry forward. It is the most important part of the

Adaptive framework. It will also track the

preferences, achievements, and activities during the

whole assessment process. According to VanLehn

and du Boulay (Vanlehn, 2006), it works as a

combination of loops, the outer loop decides which

task should be offered to the learner and the inner loop

organizes the steps to complete the task assigned to

the learner by the outer loop. This process in the

adaptive model is implemented with the help of a

learning algorithm.

3.5 The Group Learner Model

The group learner model contains all the information

regarding each group of learners such as their domain,

level of knowledge, assessment pattern.

EKM 2022 - 5th Special Session on Educational Knowledge Management

604

Moreover, on our group learner model, we focus

on both size and group composition. There is limited

research in CSCL on the effects of the size of the

group. But there is recognition that group size

depends on the scope, duration, and complexity of the

assessment activity. The group learner, however,

needs to be small enough to enable students to

participate fully and to build group cohesion (Tammy

Schellens and Martin Valcke, 2006).

Our learner group model also focuses on the

heterogeneity of groups in terms of personality traits

and performance levels. The quality of the learning

process in the context of collaborative work highly

depends on the characteristics of the group. Related

work showed the importance of personality attributes,

gender, school background, ethnic background,

motivation (Liana Razmerit, Armelle Brun, 2011) in

group performance. Another important criterion in

group composition is the learning style (Martín,

Estefanía; Paredes Barragán, Pedro, 2004). It has

been determined that the standard and quality of

learning in groups is influenced by their diversity.

Heterogeneous groups may outperform homogeneous

groups. Some studies emphasized that heterogeneous

groups may be more creative and innovative (Nijstad,

B. A., & Paulus, P. B., 2003) and they may be more

effective for individual learning.

The ability to change the group member

composition in real-time and dynamically enables the

leveling up of assessment results and improvements

in the learners' social relationships. The group learner

model is a representation of the profile of each group

of learners deduced by the collaborative assessment

activities. To illustrate the group, we need the

parameters of each group (meta-parameters) and a

grouping system that relies on grouping algorithms to

compose them.

3.6 The Assessment Model

The hypothesis put forward by the assessment model

suggests that the resolution of assessment tasks might

be justified by either the mastery or the non-mastery

of certain knowledge components (KCs) derived

from the domain model. This in its turn permits

information to be transmitted from one activity to

another. The specification of KC is essential for each

assessment activity since these components are

involved in the process of answering the instructions

of each activity by the learner.

Question & Test Interoperability (QTI) provides

a good starting point for modeling and designing

assessment systems. The Design of assessment tests

to our proposed model needs to conform to the IMS

QTI standard to better the reusability of the

assessment system and provide the basis for

interoperability specifications for the assessment

creation process: from construction to evaluation.

The QTI standard (Consortium, 2020) specifies

how to represent assessment tests and the

corresponding result reports. Figure 3 illustrates part

of a test and the way the items are structured into

sections, sub-sections, and assessment items.

Figure 3: The structure of the test Reproduced from

(Consortium, 2020).

An assessment item should not be confused with a

“question”. It is more than that since it is an

amalgamation of elements: it involves the question

and the instructions of how this question should be

introduced, as well as the answer treatment to be

applied to the candidate’s response. To present the

same question but in varied manners, the presentation

provides the structure for defining different

possibilities for the same question. Each answer

within the question can also have different structures.

The response processing determines the assessment

method. Results of a test can be recorded and saved

for future reference by other systems (Consortium,

2020).

Figure 4 illustrates the core metaclass,

AssessmentItem, in the model and how other

metaclasses are connected to it. According to

(Consortium, 2020), the ItemBody metaclass

represents the text, graphics, media objects, and

interactions that describe the item's content and

information concerning how it is structured

(Consortium, 2020).

The FeedbackBlock metaclass is very important

to present any material to the students. The feedback

that the QTI system provides is based on the result of

responseProcessing. It is controlled by the values of

outcome variables (Consortium, 2020).

To well define the technical structure of the

question and guarantee its interoperability, our

assessment model will be conform to the IMS QTI

specification. We also refer to ontology, as defined in

Test

Section

Sub‐section

Assessmentitem

Sub‐section

Assessmentitem

Section

Sub‐section

Towards an Adaptive Intelligent Assessment Framework for Collaborative Learning

605

Figure 4: A meta-classes for the assessment item.

(Gruber, 1993) as “specifications of

conceptualizations”. It is indeed a semantic

representation of complex knowledge intended for

the development of intelligent applications. It is also

defined as social constructions intended for

communication and the crystallization of domain-

specific knowledge. For this purpose, we used

ontology. The ontological model illustrated in figure

5 provides all the features used in practice while

doing assessment tests.

Figure 5: Graphical representation of the assessment

ontological model.

4 DISCUSSION

There are several varied models for representing

knowledge, teaching styles, and student knowledge.

Each model has its advantages and disadvantages. By

reviewing several works, we can conclude that

adaptation has been used in many learning contexts. A

personalized adaptive learning framework has been

constructed based on a recommendation model of the

personalized learning path and following four aspects,

namely learner profiles, competency-based

progression, personal learning, and flexible learning

environments (Hongchao Peng, Shanshan Ma,

Jonathan Michael Spector, 2019). Similarly, to

provide a method of assessing the difficulty of

learning content and students’ knowledge proficiency,

Elo-rating is a method that was designed to assist the

instructor in assessing large course programming

assignments throughout the semester (Boban Vesin,

Katerina Mangaroska, Kamil Akhuseyinoglu, Michail

Giannakos, 2022). Compared with our work, we

consider our approach adequate to be used in an online

collaborative environment. To do that, we added two

additional models to the classical architecture of an

adaptive system so that it covers adaptation on both

assessment and collaboration. The assessment model

formalized based on IMS/QTI standard and described

using ontology allows the development of the

framework using IMS/QTI specification and the

verification of the conformity of an item to the

IMS/QTI specification to guarantee its reuse and

interoperability.

5 CONCLUSION AND FUTURE

WORK

Designing an intelligent assessment framework in an

online collaborative environment presents us with a

major challenge: ensuring adaptation. While several

adaptation models exist in learning systems, this is

not yet the case for the adaptation of the assessment

strategy in an online collaborative framework. The

scope of the article is an adaptive approach for a

collaborative assessment framework in an online

environment. First, we designed a meta-model for the

collaborative assessment platform that acts as the

communication interface of our adaptive assessment

framework. Then, we proposed the extension of the

general architecture of an adaptive system which

allowed us to circumvent three important dimensions:

collaboration, assessment, and adaptation. In addition

to the domain model, the learner model, and the

adaptative model, we proposed to add two other

models to the final adaptive system: the learner group

model and the assessment model conformed to the

standardized formalism IMS/QTI. To well formalism

and describe it, we referred to ontology. A meta-

model using UML diagrams has been as well-

developed covering assessment resources content and

assessment results sections.

Finally, our perspective for this research is to

study the implementation and then the validation of

our assessment system for various assessment

domains and various learner and group profiles.

EKM 2022 - 5th Special Session on Educational Knowledge Management

606

REFERENCES

Abdalla Alameen, B. D. (2019). Adaptive eLearning

System: Conceptual Framework for Personalized Study

Environment. International Journal of Emerging

Technologies in Learning (iJET), 14(21), 28-45.

Allaymoun, H. (2014). Analysis of Discourse Threading

and Inter-animation in CSCL Chats. Proceedings of the

10th International Scientific Conference eLearning and

software for Education. Bucharest.Romania.

Allaymoun, M. H. (2021). Analysis of CSCL chats for

cognitive assessment and individual participation.

International Journal of Computing and Digital

Systems, 181-190.

Angelique, D. (2004). State of the art of interaction

Analysis: Interaction Analysis Indicators. Interaction

and Collaboration Analysis supporting Teachers’ and

Students’ Self-regulation. ICALTS.

Belcadhi, L. C. (2016). Personalized feedback for self-

assessment in lifelong learning environments based on

semantic web. Comput. Hum. Behav., 562–570.

Bhawna Dhupia, Abdalla Alameen. (2019). Adaptive

eLearning System: Conceptual Framework for

Personalized Study Environment. In A. K. al.,

Advanced Informatics for Computing Research (pp.

334–342). Singapore: Springer.

Blake, Robert J. (2009). The Use of Technology for Second

Language Distance Learning. The Modern Language

Journal, 93(s1), 822-835.

Boban Vesin, Katerina Mangaroska, Kamil

Akhuseyinoglu, Michail Giannakos. (2022). Adaptive

Assessment and Content Recommendation in Online

Programming Courses: On the Use of Elo-Rating. ACM

Trans. Comput. Educ.

Chan, T. W., & Baskin, A. B. (1990 In C. Frasson, & G.

Gauthier (Eds.)). Learning companion systems. In

Intelligent Tutoring Systems: At the crossroads of

artificial intelligence and education (pp. 6-33).

Norwood: Ablex Publishing.

Conati, C. (2009). Intelligent tutoring systems: New

challenges and directions. IJCAI, 9, 2-7.

Consortium, I. G. (2020, September 14). IMS Question and

Test Interoperability (QTI): Assessment, Section and

Item Information Model Version 3.0. Retrieved March

2022, from

www.imsglobal.org/sites/default/files/spec/qti/v3/info/

index.html#RootAttribute_AssessmentItem_qti-

context-declaration

Constantino González, Maria de los Angeles & Suthers,

Dan & Escamilla, Jose. (2003). Coaching Web-based

Collaborative Learning based on Problem Solution

Differences and Participation. I. J. Artificial

Intelligence in Education, 13, 263-299.

David.W. Johnson and Roger.T. Johnson. (1992). Positive

Interdependence: Key to Effective Cooperation.

Interaction in cooperative groups. New York:

Cambridge University Press.

Dillenbourg, P. (1999). What do you mean by collaborative

learning? In Collaborative Learning: Cognitive and

Computational Approaches. (pp. 1-19). Oxford:

Elsevier.

Gembarski, Paul. (2020). Three ways of integrating

computer-aided design and knowledge-based

engineering. Proceedings of the Design Society:

DESIGN Conference. Comput. Educ.

Gruber, T. R. (1993). A translation approach to portable

ontology specifications. Theoretical Economics

Letters, 117-243.

Hayashi, Y. (2014). Togetherness: Multiple pedagogical

conversational agents as companions in collaborative

learning. In K. E. S. Trausan-Matu, Intelligent Tutoring

Systems (pp. 114-123). Spring International Publishing.

Hongchao Peng, Shanshan Ma, Jonathan Michael Spector.

(2019, 9 6). Personalized adaptive learning: an

emerging pedagogical approach enabled by a smart

learning environment. Smart Learning Environments,

Springer.

Jennifer K. Olsen, Daniel M. Belenky, Vincent Aleven,

Nikol Rummel. (2014). Using an intelligent tutoring

system to support collaborative as well as individual

learning. Lecture Notes in Computer Science, S.

Trausan-Matu, K. E. Boyer, M. In Intelligent Tutoring

Systems (pp. 134-143). Springer International

Publishing.

Jeremy Roschelle and Stephanie D Teasley. (1995). The

construction of shared knowledge in collaborative

problem-solving. Computer-supported collaborative

learning (pp. 69– 97). Springer.

John Whiting and David. Bell. (1987). User modeling in

open learning systems. In Tutoring and monitoring

facilities for European open learning (pp. 219-237).

Amsterdam: Elsevier.

Karel Kreijns, P. A. (2003). Identifying the pitfalls for

social interaction in computer-supported collaborative

learning environments: a review of the research. In

Computers in Human Behavior (pp. 335-353).

Liana Razmerit, Armelle Brun. (2011). Collaborative

Learning in Heterogeneous Classes: Towards a Group

Formation Methodology. International Conference on

Computer Supported Education (CSEDU).

Lipponen, L. (2002). Exploring foundations for computer-

supported collaborative learning. . In Proc CSCL. 2002

(pp. 72-81).

Mariel Miller, Allyson Hadwin. (2015). Scripting and

awareness tools for regulating collaborative learning:

Changing the landscape of support in CSCL. In

Computers in Human Behavior (pp. 573-588). Spain.

Martín, Estefanía; Paredes Barragán, Pedro. (2004). Using

learning styles for dynamic group formation in adaptive

collaborative hypermedia systems. Workshop Adaptive

Hypermedia and Collaborative Web-based Systems

(AHCW´04), held in conjunction with the International

Conference on Web Engineering (ICWE´04). Munich,

Germany.

Millán, P. B. (2007). User models for adaptive hypermedia

and adaptive educational systems. Lecture Notes in

Computer Science, 4321, 54-57.

Nijstad, B. A., & Paulus, P. B. (2003). Collaborative

learning in heterogeneous classes towards a group

Towards an Adaptive Intelligent Assessment Framework for Collaborative Learning

607

formation methodology. In Group creativity:

Innovation through collaboration (pp. 326-346). New

York: Oxford University Press.

Nwana, H. S. (1990). Intelligent tutoring systems: An

overview. Artificial Intelligence Review, 4, 251-277.

Olga C. Santos, Antonio R. Anaya, Elena Gaudioso, Jesus

G. Boticario. (2003). Helping the tutor to manage a

collaborative task in a web-based learning environment.

AIED: Supplementary Proceedings. Madrid, Spain.

Pea., R. D. (1994). Seeing What We Build Together:

Distributed Multimedia Learning Environments for

Transformative Communications. The Journal of the

Learning Sciences, 285-299.

Peter Brusilovsky. (2000). Adaptive Hypermedia: From

Intelligent Tutoring Systems to Web-Based Education.

Intelligent Tutoring Systems, 5th International

Conference, (pp. 19-23). Montréal, Canada.

Rohit Kumar, Carolyn P. Rosé. (2011). Architecture for

building conversational agents that support

collaborative learning. In Learning Technologies (Vol.

4, pp. 21-34). IEEE Transactions.

Rohit Kumar, Carolyn. P. Rosé, Yi-Chia. Wang, Mahesh

Joshi, and Allen Robinson. (2007). Tutorial Dialogue as

adaptive collaborative learning support. Frontiers in

artificial intelligence and applications, 158, 383.

Soller, A. (2001). Supporting social interaction in an

intelligent collaborative learning system. International

Journal of Artificial Intelligence in Education

(IJAIED), 40-62.

Stahl, G. (2006). Group Cognition: Computer Support for

Building Collaborative Knowledge. The MIT Press.

Steven Oxman and William Wong. (2014). White paper:

adaptive learning systems. Integrated Education

Solutions.

Susan Finger, Dana Gelman, Anne Fay, Michael

Szczerban. (2006). Assessing collaborative learning in

engineering design. International Journal of

Engineering Education, 637-644.

Tammy Schellens and Martin Valcke. (2006). Fostering

knowledge construction in university students through

asynchronous discussion groups. Comput. Educ., 46,

349-370.

Tardif, J. (1999). Intégrer les nouvelles technologies de

l’information. Quel cadre pédagogique? Paris: ESF

Éditeur.

Vanlehn, K. (2006). The Behavior of Tutoring Systems.

International Journal of Artificial Intelligence in

Education, 16, 227–265.

Wenger, E. (1987). Knowledge communication systems. In

Artificial Intelligence and Tutoring Systems. Los Altos

(CA): Morgan Kaufmann Publishers.

EKM 2022 - 5th Special Session on Educational Knowledge Management

608