Using Misunderstanding and Discussion in Dialog as a Knowledge Acquisition or Enhancement Procecss - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 73

Figure 1. KB associated graphs

(a) Γ 1 associated graph (b) Γ 2 associated graph

nex is equivalent to BC connex. And according to connex graph properties: G A ∪ G B connex and G B ∪ G C
connex implies G A ∪ G B ∪ G C connex. So A ∪ B ∪ C is connex.
The agent situation we envisage is such that agents will not attempt to increase the number of their
connex components. However, there will be some cases where the 'student' will be forced to do so.
Fortunately, in some other cases, learning new knowledge may link two connex components into a new
larger one, decreasing the components total number (according to Theorem 1).

dialog: using f unctional r oles (fr )

A dialog session is the image of a lesson . A lesson is performed either because the presumed 'student' has
been asking for an important piece of information (not limited to a simple yes-or-no question or “where
do I find something...” questions), or because the 'teacher' finds him/herself in a situation where he/she
has to transmit his/her knowledge. Transposed to artificial agents situation, both cases are available. In
those, the assigned 'teacher' must know what knowledge to teach to the 'student': Therefore a lesson
has to be planed. It is then composed of several elements, each of them contained, in our framework,
in a logic formula. In our model for artificial agents, the teaching agent provides each formula to the
'student'. However, before that, the 'teacher' waits for the 'student”s understanding (or misunderstand-
ing) message of the last formula. If the 'student' doesn't understand or is not at ease, it can just inform
its interlocutor of the misunderstanding or, requests a particular information bit.
The FR theory, that models exchanges in this dialog, provides a role attachment to each utterance.
Both agents, when receiving a message, know its role and come up with an adequate answer. In our
framework, knowledge about roles is possible, because opacity, natural to human situation, is absent.
A particular clause is expressed and conveyed to the interlocutor. This clause denotes the dialogic role
of the formula to be transmitted. At the same time, this clause provides an indication about the formula
evaluation. We assign the type knowledge to universal or existential general logical formulas. This
type is neutral from the evaluation point of view, that is, knowledge might either be true or false, as a
formula. Whereas, we assign the type information to constant-uttering answer (or question), i.e. which
value is 'true', 'false', or 'unknown'. Knowledge and Information, when exchanged, might be stamped
as 'inevitably true' or 'possibly false' or, and this takes us out of the first-order logic language, 'un-
known', i.e. not evaluable to both locutors. What makes the evaluation possible, is a complex result of