Next: A First Look at the
Up: The FLUIDS Approach Towards Intelligent
Previous: Design of Structured Knowledge Models
The ability of an intelligent user interface to provide broad and
varied information related to the control task must be complemented
with an adequate presentation of the information managed by the
decision support component.
The specific requirements which imply the use of interactive
multimedia devices and intelligent multimedia techniques for this
purpose are the following:
- adaptability to a wide range of users and presentation
situations;
- more effective and appealing presentations through the
combination of different communication modes (e.g., text and
graphics) so that the strength of one medium will overcome the
weakness of another;
- the ability to react by obtaining feedback from the user about
the information presented.
The capability to flexibly tailor presentations to a user's individual
needs constitutes one of the key aspects of an intelligent multimedia
component.
Handling dynamic, non-deterministic streams of application data,
however, does not only demand for flexible design of media objects
rather than canned presentations but also requires dynamic adaption of
interaction styles.
Multimedia dialogs can be considered as a sequence of communicative acts
to achieve certain goals.
With the solution proposed in FLUIDS
, multimedia interaction is
regarded as a non-routine design task and a uniform planning approach
is employed for the knowledge-based design of multimedia dialogs.
Intelligent multimedia dialog management in FLUIDS
comprises the
following interrelated design tasks:
(1) design of the dialog structure,
(2) design of presentation structures for conveying communicative acts,
and (3) design of media objects.
Knowledge about design techniques is represented by declarative design
strategies which are treated as operators of the planning system.
Both, design strategies as well as the resulting design structures
depend on speech-act theoretic
concepts that are adapted from natural language processing to
multimedia interaction.
Inter- and intra-media coherency relationships are modeled in a
generalized version of Rhetorical-Structure Theory (RST).
Figure 3:
Plan Operator for Describing a Situation
 |
Two examples for presentation strategies are sketched in
Fig. 3 and 4.
A situation like a line-delay could be presented to the operator
by (1) informing him or her about the type of problem using a short
spoken message, and (2) describing the involved component.
A strategy to inform the operator about a cause is to direct attention
to the involved component and to describe its relevant attributes.
As these strategies illustrate, presentation planning has to employ
temporal reasoning to cope with metric and qualitative temporal
constraints for dynamic multimedia coordination.
Figure 4:
Plan Operator for Describing a Cause
 |
In most cases, the activation of the planning process is query-driven.
The dialog manager may receive information from the decision support
problem solver in reply to a basic user question or alternativly
follow-up questions related to the previous presentation have to be
treated.
Event-driven behaviour results if the intelligent interface needs to
be pro-active to issue alarms, warnings, and messages.
Taking this flexibility into account, an important advantage of the
plan-based approach to multimedia interaction becomes obvious.
Dynamic synthesis of the dialog structure enables adaptive dialog
behaviour at run-time.
Next: A First Look at the
Up: The FLUIDS Approach Towards Intelligent
Previous: Design of Structured Knowledge Models
Gerd Herzog
Last update: Sun Aug 3 18:49:55 MET DST 1997
Send comments to herzog@acm.org