e-Teaching Assistant

A Social Intelligent Platform Supporting

Teachers in the Collaborative Creation of Courses

Marco Mesiti, Stefano Valtolina, Simone Bassis, Francesco Epifania and Bruno Apolloni

Department of Computer Science, University of Milano, Via Comelico 39, Milano, Italy

Keywords:

Social Networks, Materials’ Quality and Reuse, User Reputation, Gamiﬁcation, Recommendation System.

Abstract:

With the ambition of providing teachers with a concrete tool for worldwide exploiting didactic contents to

feature their courses, we face the problem of creating a social platform with adequate functionalities to satisfy

the teacher expectations. Starting with a well designed architecture we endow it with three key functionalities

that become the stakeholders of the emerging social network: 1) a quality system ensuring the value of the

materials the users put in the platform repository as their contribution to the social business, 2) a recommender

system based on computational intelligence techniques constituting the principal tool to guide teachers along

the assembling of materials into courses, and 3) a gamiﬁcation system, root of the no-proﬁt business plan of

the platform, to involve teachers in the social network. As a result we delineate an ecosystem where teachers

exploit contents of a repository to which contribute by themselves. They are encouraged in exploiting and con-

tributing because the contents are of high quality; they are wisely assisted in the exploration of the repository

by platform services yet under their full control; and they are variously reworded by this involvement.

1 INTRODUCTION

In the last few years we have observed the prolifer-

ation of platforms (like Merlot, Connexions, Open-

Learn, ARIADNE, MACE, Share.TEC) that make

available to teachers didactic materials that can be

used for teaching. Moreover, multimedia representa-

tion models like Learning Object – LO (Wiley, 2000),

Open Educational Resources – OER (Atkins et al.,

2007), and SCORM (ADL - Advances Distributed

Learning, 2004), have been proposed for enhancing

the interoperability of platforms in representing and

exchanging didactic resources. By means of these

platforms/models and also the materials made avail-

able on the Web by Schools and Universities, a huge

amount of didactic materials is available that could be

adopted (or acquired when subjected to fees) for the

preparation of single lessons or entire courses. In this

overwhelming of information, however, it is not easy

to discover the right materials that meet the prepara-

tion and expectation of the class students. Moreover,

the quality of the resources is not always the same,

the level of detail of the treated topics ranges differ-

ently from elementary to very detailed and advanced

presentations, and the requirements for effectively at-

tending to the materials are not always clear. There-

fore, teachers wishing to reuse already developed ma-

terials spend hours in the retrieval of adequate lec-

tures, exercises and projects for their classes, succeed-

ing only seldomly in ﬁnding the right ones.

In this paper we wish to detail the characteristics

of a tool, named e-Teaching Assistant, speciﬁcally tai-

lored for helping and supporting teachers in the pro-

cess of preparation and sharing of didactic resources

(either single materials or entire courses). This tool

is designed by considering teachers as “demanding-

users”, namely individuals that, accustomed to pro-

duce educational materials and having clear ideas on

the topics to be taught according to the level of prepa-

ration of the class to which they are intended, do not

expect to receive “pre-deﬁned” instructions on how to

create and organize their courses; rather, they demand

to both interact with the system and discuss and col-

laborate with domain experts. A key characteristic of

e-Teaching Assistant is that users can formally or in-

formally collaborate for the realization of courses by

means of a social network (SN) speciﬁcally tailored

for this context. Moreover, intelligent services, de-

noted as “meta-services”, are devised for our demand-

ing users, supporting them in the preparation, retrieval

and exchange of materials. Among them, we point

out: the reviewing service for improving the quality

569

Mesiti M., Valtolina S., Bassis S., Epifania F. and Apolloni B..

e-Teaching Assistant - A Social Intelligent Platform Supporting Teachers in the Collaborative Creation of Courses.

DOI: 10.5220/0004963505690575

In Proceedings of the 6th International Conference on Computer Supported Education (CSEDU-2014), pages 569-575

ISBN: 978-989-758-020-8

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 1: e-Teaching Assistant Architecture.

of the published materials by taking into account the

characteristics of the students to which they are de-

voted (language, age, student backgrounds, and ob-

jectives); the recommendation service for the identi-

ﬁcation of suitable materials; composition services to

generate new courses by reusing materials developed

by other colleagues, still keeping the provenance of

the materials; gamiﬁcation services for encouraging

teachers in proactively participating to the platform.

This paper presents our thoughts on how these

meta-services should be integrated to forecast a social

intelligent learning management system for demand-

ing users. We start with an overview of the whole

architecture, giving prominence to the enhancements

w.r.t. state-of-the-art e-learning platforms. Then, we

highlight the features of the proposed recommenda-

tion and gamiﬁcation systems, in order to face the dis-

tinctive peculiarity of the afforded task.

We are currently experimenting this new paradigm

in the framework of the NETT European project (nett-

project.eu), where these services aim to help teachers

to improve the provision of courses for improving the

entrepreneurship (Valtolina et al., 2014). This disci-

pline is crucial in the education of young generations,

yet not well assessed as a corpus of basic lessons and

those orienting the learner to a given specialization.

We face this immaturity by a cognitive recommenda-

tion system, as for advanced service, and a gamiﬁ-

cation system, as for gathering community members

and their evaluations at the basis of the former. In this

short note we will discuss these tools from a method-

ological perspective, having our experience in NETT

as a workbench.

2 ARCHITECTURE AND

QUALITY OF THE MATERIALS

In recent years many e-learning platforms have been

devised, like Merlot, Connexions, OpenLearn, ARI-

ADNE, MACE and Share.TEC, mainly focused on

handling single resources (like powerpoint presenta-

tions, pdf ﬁles, exercises and so on). In Merlot and

Share.TEC the architecture relies on the deﬁnition of

ontological structures to support the sharing of digital

content. In particular, Share.TEC proposes an ontol-

ogy called TEO (Teacher Education Ontology) (Ivino

et al., 2009) to provide a powerful tool for cataloging

and classifying materials capable to provide person-

alized access to didactic resources, based on the ac-

tual users needs, their cultural context, and their pro-

fessional proﬁle. Even though these systems provide

some SN capabilities, they are quite limited and not

well integrated in the entire process of production and

use of the developed resources.

The architecture to be devised for e-Teaching As-

sistant should meet the following requirements:

• Didactic materials should be handled at differ-

ent levels of aggregation, ranging from single re-

sources to modules and courses, where a module

is an aggregation of single resources and a course

CSEDU2014-6thInternationalConferenceonComputerSupportedEducation

570

is an aggregation of modules. In this way, teach-

ers can have a more complete overview of how the

course and other ancillary materials (like articles,

exercises, discussion) associated to a module or an

entire course are organized by other colleagues.

• Each course, module or single resource is associ-

ated with a set of metadata that describes the di-

dactic materials and give meaning to them. By

starting from the standard LOM (IEEE, 2006) we

consider a small subset that aptly conform to the

context under consideration. Note that: the meta-

data for modules (and in turn those for courses)

can be automatically extracted from the annexed

resources; and, multiple values can be speciﬁed

for the same property (e.g. the language of the

module is the union of those used in its contents).

• Exploiting the result of the Merlot and Share.TEC

projects, we extend and integrate our metadata

structure using the OAI-PMH protocol for the se-

lective gathering of metadata describing learning

objects

. Through this protocol, our metadata

structure is extended in an ontology able to allow

personalized access to the educational materials

and to offer an effective strategy for integrating

different didactic content sources.

• The didactic materials are not forced to be stored

within e-Teaching Assistant. Materials can be

present in other platforms or made available

through web services. However, their metadata

are locally stored and exploited by the metaser-

vices. External materials might also be subjected

to the payment of royalties to their authors or to

the platforms where they are stored.

• A SN should be deeply integrated in the system

in order to offer social metaservices for the cre-

ation of communities around the topics covered

by e-Teaching Assistant and the support of peers

in all the phases of the creation, revision, audit

and publication of didactic materials as well. The

actors of the SN are classiﬁed in Visitors, Contrib-

utors, Masters (leaders in given topics), and Ex-

perts (contributors with large experiences). These

roles will dynamically change, according to the

level of participation to the network.

• Levels of reputations of the SN members, levels of

appreciations of the developed materials, prove-

nance of the developed materials, and their reuse

will be associated to the actors and materials han-

OAI-PMH is the Open Archives Initiative Protocol for

Metadata Harvesting (http://www.openarchives.org/pmh/)

whose aim is to create an independent interoperability

framework based on metadata harvesting.

dled by the system, maintained up-to-date and ex-

ploited by the available metaservices as well.

At any level of the education system, an e-learning

platform needs to face issues concerning the quality of

didactic materials offered to students. In Merlot a re-

viewing system similar to the one adopted in the con-

text of publication of journal papers is used. More-

over, all the cited platforms support the grading of the

materials using different scaling (either from 0 to 5

or from 0 to 10): an information that is only used for

ranking the materials when they are retrieved.

In e-Teaching Assistant we will adopt the afore-

mentioned solutions to guarantee the quality of the

developed materials, enhancing them by:

• integrating the reviewing activities with com-

ments and ideas coming from the SN. The SN

should thus become a mean for exchanging ideas,

comments and solutions for better facing the

learning issues of students. Both formal and in-

formal communications will be granted to teach-

ers belonging to the same community;

• linking the quality of the developed materials

with the respectability of the teachers producing

them. In this way, teachers are encouraged to pro-

duce high quality materials to improve their re-

spectability in the community. The use of levels

of respectability has also the advantage of identi-

fying masters of given topics and experts that can

help the former in the reviewing processes.

3 COMPOSITION OF COURSES

The core business of e-Teaching Assistant is the com-

position of new courses. To this aim teachers are

guided to both organize courses as a wise sequence

of modules and to ﬁll up modules with resources that

comply with their didactic goals and cultural pref-

erences. Thus, besides the traditional tools for re-

trieving didactic materials based on keywords and

metadata matchings, the intelligence of e-Teaching

Assistant is represented by a computationally intelli-

gent recommendation system based on both metadata

and user consensuses variously collected through the

common social tools of the SN.

Actually, Recommendation System (RS) is a rel-

evant component of every modern SN in most dis-

parate ﬁelds, ranging from movies, music, books, to

ﬁnancial services. Usually implemented as a web ap-

plication, it constitutes a class of algorithms aimed

at predicting user responses to options, by generating

meaningful recommendations to a collection of users

for items that might interest them.

e-TeachingAssistant-ASocialIntelligentPlatformSupportingTeachersintheCollaborativeCreationofCourses

571

In e-Teaching Assistant, RS will focus primarily

on the exploitation of didactic materials available in

already developed platforms that take into account the

characteristics of both teachers and students in terms

of backgrounds, competences, language, and level of

instruction. Moreover, in order to recommend materi-

als with high quality, it will exploit social aspects such

as the interaction between teachers and students, their

reputation, and the respectability they have gained in

the network. Such social context well embraces the

proliferation of open educational resources (OERs)

released under Creative Commons licenses, and the

variability of the characteristics of the community of

users involved in the educational process as well. On

the one hand the heterogeneity of the on-line mate-

rial, developed in different languages and with differ-

ent quality for a variety of target students in differ-

ent contexts of learning, complicates the realization

of useful RSs. On the other hand, teachers are rarely

satisﬁed by predeﬁned and non-ﬂexible recommenda-

tions when they exploit materials developed by other

colleagues. Moreover, students having different back-

ground, culture and level of instruction need speciﬁc

recommendations to support the discovery of suitable

materials according to their aims, interests, and di-

dactic needs. Finally, both teachers and students are

tightly connected by means of SNs through which

they can chat, exchange materials, and give evalua-

tions on the resources available in a recommendation.

Traditionally, RS are classiﬁed as:

• Collaborative Filtering (CF) Systems, where pre-

dictions about the interests of a user are inferred

on the basis of people having similar interests

and preferences. They are based on k-nearest

neighbor (kNN) methods (Breese et al., 1998) as

for Neighborhood-based approach and paramet-

ric estimation techniques as for Model-based ap-

proach (Bell et al., 2009).

• Content-Based (CB) Systems, where recommen-

dations rely on the user’s preference and the

items’ descriptions. They are based on query

and relevance/similarity scores as for IR ap-

proaches (Mooney and Roy, 2000), and on clas-

siﬁers of hcontent, user-ratingi pairs, such as

Naive Bayes (Pazzani and Billsus, 1997) and

kNN classiﬁers, decision trees, and neural net-

works (Melville et al., 2002) as for Classiﬁer.

• Hybrid Recommendation Systems, combining the

above approaches in order to mitigate the asso-

ciated limitations, for instance via boosting tech-

niques (Schein et al., 2002) or generative mod-

els (Kim and Ahn, 2012).

Recent studies have attempted to use techniques for

studying social relationships in terms of network the-

ory, the Social Network Analysis (SNA), in com-

bination with RSs. In (Brusilovsky, 1996) the au-

thors got encouraging results by assigning weights

to the content-based attributes used for recommenda-

tions as a function of their importance for users. In

the e-learning context, several RSs have been devel-

oped to propose courses, materials, and relevant top-

ics in forums (Brusilovsky, 2012). Although increas-

ingly popular, so far only few studies such as (Frias-

Martinez et al., 2006; Mulwa et al., 2010) have been

addressed to suggest collaborative learning resources.

These methods are not sufﬁcient to make satisfac-

tory suggestions, mainly for the following reasons:

1. The inability to treat the uncertainty of both the

ratings/suggestions and the resource description

proposed by the SN members. In fact, judgments

collected from a plethora of users with different

habits and cultures may produce contradictions

which in turn result in data having a high degree

of ambiguity. This calls for a granular interpreta-

tion of the information provided by such crisp at-

tributes, for instance in terms of fuzzy sets, rough

and interval sets, and so on (Apolloni et al., 2008).

2. The lack of interpretability of the recommenda-

tion model and, as a consequence, of the recom-

mendation policy.

3. The lack of user interaction. With the advance-

ment of computational techniques, we have the

unprecedented ability to allow machines to assist

users in completing their tasks. Thus, fully auto-

matic suggestions may not be entirely appreciated

by the active user, which risks to get frustrated in

using the RS platform whenever its recommenda-

tions prove to be erroneous.

4. The inability of current RSs to highlight the social

relationships between users. This is especially im-

portant within a SN where each user will certainly

appreciate receiving recommendations from those

considered “closest” to her (classmates, teachers,

etc.) (Shinha and Swearingen, 2001).

We expect e-Teaching Assistant to overcome these

limitations by embedding computational intelligence

into a hybrid system through the following features:

i) a Rule-Based System (RBS), to provide the user

with an interpretable recommendation policy. We

focus on a special instance of a decision tree algo-

rithm, exploiting the one-to-one correspondence

between decision trees and RBSs.

ii) Granular Computing techniques (Apolloni et al.,

2008), which are essential to handle non crisp

judgments. By fuzzifying the sets constituting an-

CSEDU2014-6thInternationalConferenceonComputerSupportedEducation

572

tecedents and consequents of the RBS, we will in-

troduce a fuzzy reasoning based on fuzzy entropic

criteria and expressed in terms of fuzzy decision

trees. To avoid injecting inconsistencies in the ﬁ-

nal rule set, the tree construction will be compli-

ant with the multivalued attributes characterizing

modules and contents of our repository.

iii) Interactive Machine Learning techniques, to allow

the user to interactively participate in the devel-

opment of the ﬁnal recommendation. The devel-

opment of a RS guided by the user intervention

translates in dynamic modiﬁcations, backtracking

included, of the rule set on the basis of the user

input. In turn, the system will provide the most

valuable suggestions satisfying the (explicit and

implicit) constraints introduced by the user, such

as presence of introductory courses, constraint on

the course duration, and so on.

iv) Social Network Analysis (SNA), to capture so-

cial relationships among users. We will work on

ad hoc clustering techniques for ﬁnding groups

within the SN members. In this way, we will re-

place the concept of proximity of the typical CF

algorithm with the corresponding SNA one, intro-

ducing weights on the preferences of the users, so

that well-reputed members of the SN will have a

higher inﬂuence in the whole process.

4 TEACHERS’ INVOLVEMENT

STRATEGIES

The SN life is made up of social activities, by deﬁ-

nition. This implies that a SN may survive the ini-

tial enthusiastic period where the network is designed

and its mockup is implemented only if the platform

is populated by a community of users who have con-

crete motivations for interacting and sharing knowl-

edge. Having decided for an open platform that is

not rooted on a proﬁt business model, we identiﬁed

gamiﬁcation as a relevant tool for fostering the user

interest. Gamiﬁcation indeed is a familiar context in

the teaching framework for two reasons:

1. Students are in a continuous competition as a nat-

ural status of their job. It is a competition which is

primarily toward themselves: they try improving

themselves everyday, hence compete and over-

whelm their own abilities. Actually several types

of games are used in classroom activities (eLearn

Magazine Staff and Contributors, 2011; Muntean,

2011; Raymer, 2011) based on typical game el-

ements like time, accuracy, point systems inte-

grated into training programs.

2. Teachers are in continuous competition (Nah

et al., 2013) toward three frontlines: 1) themselves

as former students, 2) students, to whom they can

never yield, and 3) colleagues for both immaterial

(pride) and material (carrier) reasons.

We plan leveraging on this competition for engag-

ing teachers in creating and sharing materials. This

type of metaservice is designed as an incentive sys-

tem based on a set of rules that encourage teachers

to explore and learn the properties of their possibility

space. We adapt the common strategies (Deterding

et al., 2011) to the above competition line in terms of:

• Frontline n

◦

1. A layering mechanism which allows teachers

to learn new skills incrementally, and then

practice those skills before demonstrating their

mastery in creating new materials. Hence they

are incrementally challenged to featuring con-

tents, modules and ﬁnally entire courses.

2. A character upgrade scenario which provides

feedback to teachers for warning about how

much progress they have made in creating a

course. They gather virtual goods and assets

to change the character in the way they like.

• Frontline n

◦

1. Private or closed community groups, which

provide their agreement on the material pro-

duced by teachers according to their acquired

competencies and rules. In fact, the over-

whelming success and inﬂuence of social me-

dia in modern society corroborates the power

of other people’s opinion.

2. Objective indexes such as number of downloads

of the single contents by students, cumulated

scores expressed by them, etc, which are a di-

rect way of acknowledging the teachers work.

• Frontline n

◦

1. Keeping the authorship of the developed mate-

rials and the acknowledgment of the work done,

that is very relevant for teachers. A contrib-

utor can integrate a module developed by an-

other teacher and decide to keep it “as is” or to

modify it (by adding or removing content). The

system keeps track of the fact that the module is

duplicated from an existing one, and the com-

pliance (or not) with the original form. This

feature is useful for maintaining the provenance

of the material and to ensure the author’ royal-

ties and for incrementing his/her reputation.

2. Making teachers talking to one another, which

gives them common goals and rewards, espe-

cially if that reward is predicated on group par-

e-TeachingAssistant-ASocialIntelligentPlatformSupportingTeachersintheCollaborativeCreationofCourses

573

ticipation. Teacher’s peers see when they col-

lect these rewards. These extrinsic rewards are

much more effective if people can use them for

bragging rights, rather than just having some

extra trophy graphic that nobody else will see.

A ﬁnal remark on the game design concerns the

whole presentation of the metaservices. The de-

sign, the look and feel, the interaction style and the

communication process of the e-learning environment

need a speciﬁc care and an incremental production of

mock-ups anytime that new users requirements ap-

pear. Moreover, it is import to test our prototype as

early as possible. One of the most repeated mistake

is to make assumptions about how the target audience

will use the product. The only way the designers can

understand it is to put the system in front of them,

watch them use it, and to document the experience in

order to pay attention to how long it takes to make the

correct input, and to watch through teachers’ eyes for

seeing where they look ﬁrst on the screen mockups.

5 CONCLUDING REMARKS

The main goal of the e-Teaching Assistant is to of-

fer a new opportunity for supporting teachers by ex-

ploiting the contributions of a SN able to enhance

and enrich didactic contents proposed by their mem-

bers. To this aim, the paper proposes a social oriented

solution based on three metaservices for exchang-

ing high quality didactic materials, retrieving content

through a computationally intelligent recommenda-

tion service and stimulating the teachers involvement

through gamiﬁcation strategies. Other metaservices

are under design for offering a semi-automatic combi-

nation of modules according to the requirements and

skills characterizing them and by ﬁtting the teachers’

expectation according to their proﬁle and background.

REFERENCES

ADL - Advances Distributed Learning (2004). Sharable

content object reference model (SCORM). Overview.

http://www.adlnet.gov/About/alexandria/default.aspx.

Apolloni, B., Bassis, S., Malchiodi, D., and Pedrycz, W.

(2008). The Puzzle of Granular Computing, vol-

ume 138 of Studies in Computational Intelligence.

Springer Verlag.

Atkins, D. E., Brown, J., and Hammonds, A. (2007). A

review of the open educational resources (OER)

movement: achivements, challenges, and new

opportunities. www.hewlett.org/uploads/ﬁles/

ReviewoftheOERMovement.pdf.

Bell, R., Koren, Y., and Volinsky, C. (2009). IEEE Com-

puter. Matrix factorization techniques for recom-

mender systems, 42(8):30–37.

Breese, J. S., Heckerman, D., and Kadie, C. (1998). Empir-

ical analysis of predictive algorithms for collaborative

ﬁltering. In Proc. of the 14th Conf. on Uncertainty in

Artiﬁcial Intelligence, Madison, WI, July 1998.

Brusilovsky, P. (1996). Methods and techniques of adaptive

hypermedia. User modeling and user-adapted inter-

action, 6(2):87–129.

Brusilovsky, P. (2012). Adaptive hypermedia for education

and training. Adaptive Technologies for Training and

Education, 46.

Deterding, S., Sicart, M., Nacke, L., O’Hara, K., and Dixon,

D. (2011). Gamiﬁcation. using game-design elements

in non-gaming contexts. In CHI ’11 Extended Ab-

stracts on Human Factors in Computing Systems, CHI

EA ’11, pages 2425–2428, New York, USA. ACM.

eLearn Magazine Staff and Contributors (2011). Predictions

for e-learning in 2011. eLearn, 2011(1).

Frias-Martinez, E., Chen, S. Y., and Liu, X. (2006). Survey

of data mining approaches to user modeling for adap-

tive hypermedia. IEEE Trans. on Systems, Man, and

Cybernetics, Part C, 36(6):734–749.

IEEE (2006). Standard for Learning Object Metadata

(LOM): Conceptual Schema IEEE 1484.12.1.

Ivino, S., Bocconi, S., Boytchev, P., Earp, J., and Sarti, L.

(2009). Sharing digital resources in teacher education:

an ontology-based approach. In Demetra EOOD.

Kim, K. J. and Ahn, H. (2012). Hybrid recommender sys-

tems using social network analysis. World Academy

of Science Engineering and Technology, 64.

Melville, P., Mooney, R. J., and Nagarajan, R. (2002).

Content-boosted collaborative ﬁltering for improved

recommendations. In Proc. of the 18th National Conf.

on Artiﬁcial Intelligence (AAAI-02), pages 187–192.

Mooney, R. J. and Roy, L. (2000). Content-based book rec-

ommending using learning for text categorization. In

Proc. of Digital Libraries Conf., pages 195–204.

Mulwa, C., Lawless, S., Sharp, M., Arnedillo-Sanchez, I.,

and Wade, V. (2010). Adaptive educational hyperme-

dia systems in technology enhanced learning: a lit-

erature review. In Proc. of the 2010 ACM Conf. on

Information technology education, pages 73–84.

Muntean, C. I. (2011). Raising engagement in e-learning

through gamiﬁcation. In The 6th International Conf.

on Virtual Learning ICVL 2011, pages 323–329.

Nah, F. F.-H., Telaprolu, V. R., Rallapalli, S., and Venkata,

P. R. (2013). Gamiﬁcation of education using com-

puter games. In HCI (15), pages 99–107.

Pazzani, M. J. and Billsus, D. (1997). Learning and revis-

ing user proﬁles: The identiﬁcation of interesting web

sites. Machine Learning, 27(3):313–331.

Raymer, R. (2011). Gamiﬁcation: Using game mechanics

to enhance elearning. eLearn Magazine, 2011(9):3.

Schein, A., Popescul, A., Ungar, L. H., and Pennock, D. M.

(2002). Methods and metrics for cold-start recommen-

dations. In Proc. of the 25th annual Intern. Conf. on

Research and Development in Information Retrieval

(SIGIR), pages 253–260, NY, USA, 2002. ACM.

CSEDU2014-6thInternationalConferenceonComputerSupportedEducation

574

Shinha, R. and Swearingen, K. (2001). Comparing recom-

mendations made by online systems and friends. In

Workshop on Personalization and Recommender Sys-

tems in Digital Libraries, Dublin.

Valtolina, S., Mesiti, M., Epifania, F., and Apolloni, B.

(2014). Towards a social e-learning platform for de-

manding users. In IEEE EDUCON. To appear.

Wiley, D. A. (2000). Connecting learning objects to instruc-

tional design theory: A deﬁnition, a metaphor, and a

taxonomy. The Instructional Use of Learning Objects:

Online Version.

e-TeachingAssistant-ASocialIntelligentPlatformSupportingTeachersintheCollaborativeCreationofCourses

575