Information Technology Reference
In-Depth Information
lecting data such as amount of time spent, links
on pages visited, scrolling activity, test results
and page navigation history. Navigation history
is defined by a Markov model implemented using
a finite order Markov chain. A finite number of
profiles describing the user sophistication level
are assumed. Using the aforementioned data and
Bayesian classification, probability that the user
belongs to any of the profiles is determined.
The system provides adaptive navigation and
presentation. Presented learning pages are cre-
ated dynamically by selecting appropriate XML
fragments. Furthermore, the same content can be
presented in different styles based on user needs.
The system includes authoring tools to convert
existing educational content to XML fragments.
remote explanation database by the moderator
agent. Explanation usefulness data is acquired
from learners. If no suitable explanation can be
retrieved from either database, the dialog agent
searches for an explanation by communicating to a
number of negotiating agents. Negotiating agents
include concept vector translation tables to associ-
ate different curriculum systems. Each negotiating
agent accesses its associated explanation databases
to retrieve an explanation. If no suitable explana-
tions can be retrieved from associated databases,
the dialog agent requests the negotiating agents to
find a user who has overcome a similar learning
difficulty in the past. Each negotiating agent has
access to a database of users indicating their skill
level (i.e., curriculum and knowledge concepts
mastered). The retrieved user communicates in
synchronous mode with the learner to assist in
surpassing the learning difficulty. At the end of
an synchronous session, the pedagogical agent
queries the learner to author the given explanation
and its usefulness. Corresponding explanation data
is sent to a moderator agent for insertion into an
explanation database.
Patents Involving Agent-Based IESs
In (Frasson & Gouarderes, 2002) an approach to
a Web-based IES consisting of multiple collabo-
rating agents is presented. The approach provides
learner curriculum data and assistance in the form
of asynchronous/synchronous explanations. It
comprises pedagogical, dialog, service, moderator
and negotiating agents. The pedagogical and dialog
agents are close to the learner, whereas the other
agents are remotely positioned. The pedagogical
agent selects the curriculum to be taught and sends
a request to the service agent for retrieving it from
a curriculum database and transferring it to the
pedagogical agent. The pedagogical agent also
selects a learning strategy. The pedagogical agent
provides the learner explanations when needed/
requested for specific knowledge concepts or
curriculum items. When required, the pedagogi-
cal agent requests explanations from the dialog
agent which handles the process of acquiring
the best explanation. Explanations can initially
be acquired from a local explanation database
or from a remote explanation database stored at
the service agent. Explanations in the databases
are indexed and ranked according to usefulness
in helping learners. This task is performed in the
Patents Involving IES
Shells/Generators
Goodkovsky (2004) presents an approach to a
generic and reusable ITS shell. The approach
provides a fuzzy domain/learner/tutor theoretical
model-based shell with dynamic planning capabil-
ity, independent of training paradigms, domains
and learners. A fuzzy graph represents domain
knowledge structure. Data and metadata can be
easily authored. Different instruction modes are
provided. The pedagogical module provides,
among others, user behavior recognition, cogni-
tive interpretation of learner actions, decisions
on sufficiency of different instruction modes and
best suitable mode to be applied next, planning
and generation of training actions. Fuzzy logic is
used to dynamically adapt the available sequence
of training actions.
Search WWH ::
Custom Search