Biomedical Engineering Reference
In-Depth Information
information state respectively if necessary. As a result, a unified dialogue model is
constructed as shown in Fig. 2 (left). Four update rules are added, so that the discourse
context can always be considered and updated; e.g. the update rule ACCEPT adds a
new assertion a into B's belief and refer it as known from then on.
Finally, we define a unified dialogue model as a deterministic recursive transition
network built at the illocutionary level of interaction processes; its transitions can
only be triggered by fulfilled conditions concerning the information state, and with
the consequences of information state update according to a set of update rules.
Fig. 2. A simple unified dialogue model and its CSP specification
3.2
A Formal Language Based Development Toolkit for Dialogue Modelling
Deterministic recursive transition networks can be illustrated as a typical finite state
transition diagram (cf. Fig. 2 left), which provides the possibility of specifying the
described illocutionary structure with mathematically well-founded formal methods,
e.g., with Communicating Sequential Processes (CSP) in the formal methods commu-
nity of computer science.
CSP can not only be used to specify finite state automata structured patterns with
abstract, yet highly readable and easily maintainable logic formalization (cf. [24]), but
it is also supported by well-established model checkers to verify the concurrent as-
pects and increase the tractability (cf. [25]). Thus, CSP is used to specify and verify
the unified dialogue models (cf. the example in Fig. 2 (right)).
To support the development of unified dialogue models within interactive systems,
we provided the Formal Dialogue Development Toolkit (FormDia cf. Fig. 3).
Fig. 3. the Structure of the FormDia Toolkit (cf. [26])
To develop the unified dialogue model based discourse management, FormDia
toolkit can be used according to the following steps:
Search WWH ::




Custom Search