From 1997, a join cooperation with the Education Science
Department of University of Roma TRE has been created to the
study the possibility of applying the information technologies
to the distance education problems.
In particular the group has participated to the European Union
Leonardo Program "Multimedia WWW Kernel For Distance
Learning" (1997-2000) and "From e-learning to m-learning"
(2001-2002). The international group was composed by University
of Roma TRE, Ericsson Ireland, and the Norwegian Distance
Learning Institute NKI Fjernundervisningen.
Current research is focused on algorithms and processing
procedures to personalize and to adapt both the message and
the learning strategies based on the background, the ability,
and the competences of the learner as well as the impact of
the learning styles, and the implementation and integration
of the afore mentioned solutions in a Learning Management
System (LMS).
In principle, two levels of adaptation can be distinguished:
the overall adaptation of the learning process ( macroadaptivity),
and the adaptation of the elementary portions that constitutes
the learning material (microadaptivity).
To introduce adaptivity to individual learners, the total
didactic message is first subdivided in elementary parts,
each characterized by its own goals in terms of acquisition
of knowledge, competences and ability.
Every elementary unit is implemented by a set of didactic
fragments which share the same goals, but use different modalities
of presentation or different pedagogical activities and that
are matched to different learning styles (with which the student
acquires and elaborates the message of instruction) and to
different learning strategies (the learning strategies comprise
both the cognitive strategies that refers to the acquisition,
retention and retrieval of different types of knowledge, as
well as the metacognitive strategies that refer to the methodologies
in order to learn how to learn).
The macroadaptivity is achieved by combining the didactic
fragments matched to the actual learner profile. Then, the
microadaptativity is obtained by modulating the message associated
to each single didactic fragment, originally formulated for
a virtual receiver (that is a receiver that possess all the
lexicon necessary in order to comprise the message) in function
of the receiving degree of verbal competences of a real receiver.
The final goal is the realization of a system that is able:
1. to recognize the elements that characterize the single
student from the point of view of the knowledge, competences
and skill and his/her preferences in relation to the learning
style (estimator of the student profile)
2. to estimate the pedagogical implications of the student
profile and to dynamically generate the personalized learning
path starting from the characteristics of single didactic
fragments;
3. to modulate the relative message of instruction of single
didactic fragments
4. to dynamically update the profile of the student in its
complex.
As preliminary step, we will explore the possibility to employ
the formalism and the theoretical body associated to the Bayesian
networks and to the fuzzy logic in order to describe and identify
the profile of the student in a multidimensional space that
comprises the subsets related to the knowledge, to the competences
(included the verbal ones), to the skill and the learning
styles. Such formalism has been employed until now to drive
the macroadaptivity of the learning path.
Actually the study of algorithms and techniques is focused
on single didactic fragments constituted by a hypertext in
HTML format, so that the modulation will consist in the expansion
of the message formulated for a virtual student into text
segments that depends on the student profile and his/her capabilities
in understanding the Italian language.
Subsequently the techniques that concur to extend the message
modulation to the cases in which the single didactic fragments
are realized by means of more complex multimedial documents
will be investigated. In particular it will be analyzed the
case in which to a video containing diagrams, animations,
movies, etc, is associated a comment carried out from the
voice of a narrator.
Once the comment text has been synthesized with the same
modalities employed for the simpler fragments, the audio will
be automatically generated by means of an out-of-shelf text-to-speech
engine. For the synchronization between the different media
that compose the fragment the modulator will automatically
produce a script in the SMIL language, recently introduced
in the framework of the multimedia communications in order
to allow the creation of hypermedia presentations composed
by synchronized texts, images, audios and videos, without
having to resort to sophisticated tools.
The possibility to integrate in the video the character of
the narrator is tied to the maturation of the technologies
for the animation of three-dimensional human models (Avatar),
incorporated in the more recent standards of multimedia coding,
whose development is not pertinent with the present research.
- MIUR FIRB project 2005-2008, "Personalization and
adaptation of didactic message and learning strategies for
e-learning applications in economics" (Procedure individualizzate
a modulazione di messaggio per la formazione e-learning
in campo economico), as local coordinator.
