Social, Personalized Lifelong Learning - E-Infrastructures and Technologies for Lifelong Learning - page 121

etc., are used. Interestingly, the paper reports
that, whilst the students might have perceived
the adaptive comments as intrusions, the overall
result (in terms of learning) was positive. Our
approach is closer to this study, as the collabora-
tive adaptation process aims at guiding students
towards useful interactions with each other, and
with their teachers (recommended learners), as
well as guiding students towards useful recom-
mended learning content based on their profiles.
However, our research also blends not only the
learning process and the collaboration process,
but also the learning and authoring processes.
Awerbuch et al. (2005) have taken an AI-driven
approach, and describe processes of adaptive
collaboration in peer-to-peer systems in terms
of players (or agents) with shared or exclusive
goals, thus cooperating or competing against
each other. Their system is not directly aimed at
learning, and its focus in on how to minimize the
cost for an agent in a world of threats (e.g., from
dishonest 'players'). Whilst this work may be
useful for collaborative and competitive systems
in general, it is less applicable in the context of
learning, where learners might try to 'beat the
system', but would usually gain little from being
dishonest to each other. Our aim is to define a new
social personalized adaptation model that can cur-
rently be applied in extant learning management
systems (LMS), in which learners and teachers
can engage in a multi-role, personalized, adaptive
learning environment to enhance the learning and
authoring processes.
In the context of lifelong learning, the
APOSDLE (Advanced Process-Oriented Self-
Directed Learning environment) project (Lind-
staedt & Mayer, 2006) introduced new ways to
support informal learning activities (work, learn,
collaborate) for the workers in their working
environments, which gives learners support ,
by providing the learners with support for self-
directed searching and learning within the work-
ing environment; experts support , by allowing
social interaction between learners, and making

the results of this interaction available to other
learners in their own learning environments; and
worker support , in which the learning process hap-
pens within the working context, and the learners
access the learning content without the need to
change the working environment. Our approach
is slightly similar as it supports recommendations
techniques, such as recommended learning content
based on the learner's profile and recommended
expert learners also based on the learner's profile.
The differences appear in the target — we target
not just workers, but lifelong learners, as well as
students in formal education.
The Ensemble (Semantic Technologies for
the Enhancement of Case Based Learning)
project (Carmichael et al., 2009) is a relatively
new project, which explores the benefits of the
Semantic Web to support learners and teachers
in a case-based learning approach. The goal of
this work is to explore both the nature and role of
the learning cases between learners and teachers,
using the emerging semantic technologies. This
work is very interesting, but is still in progress.
Our approach does not rely heavily on semantic
web techniques currently, although import from
RDF, for instance, is possible. The result of the
Ensemble project could be extended in the near
future, based on the framework we are proposing.
In the ALS project 4 , the adaptation to col-
laboration approach is similar to our approach;
however our approach extends the collaboration
by using the Social Web techniques (such as rat-
ing, tagging, etc.).
Telme (Sumi & Nishida, 2001) is a communica-
tion tool that acts as a moderator between people
with different levels of knowledge. The personal-
izationin Telme occurs by presenting information
from a knowledge base customized according to
the user's profile. The system is effective when the
user cannot question others directly by concluding
the context of the conversation from predefined
conceptual spaces.
The personalization in the work of Pinheiro et
al. (2008) is based on the mobile user's profile,