Intelligent Systems
The development of Intelligent Systems is a core AIAI activity.
Intelligent Systems are computer programs constructed using AI
technology or concepts that perform tasks requiring intelligence or
expertise. Intelligent Systems have been developed to address
diagnostic, monitoring, classification, planning, scheduling, advisory
and design tasks. In recent years, for example, AIAI has been involved
in projects to build systems to plan and schedule spacecraft assembly,
to design the layout of telephone directory pages, to detect possible
financial fraud and to help diabetics monitor their current physical
state.
Methodology
Intelligent Systems can perform a variety of tasks in a wide range of domains.
Consequently, there is no single 'Intelligent Systems architecture': the form and
content of an Intelligent System will depend on the nature of task in
question; on the source and characteristics of the knowledge to be
embodied within the system; and on the sort of environment in which the
system will have to operate. But, there is a common underlying
methodology in the construction of an Intelligent System:
task analysis, knowledge source
identification and conceptual design of solution system;
knowledge acquisition and modelling;
detailed design of system components;
implementation and initial testing;
integration and deployment;
system testing.
For certain tasks, rather than risk full-scale system development
immediately, it may be appropriate to develop a small-scale proof-of-concept prototype at an
early stage of this process to validate assumptions about the domain and
test the proposed solution principles.
AIAI can play a role in Intelligent System development projects during
any of these phases, but most typically would be involved during the
earlier stages of analysis, modelling and design, and
proof-of-concept implementation where AIAI's experience of working with
specifically AI concepts and technologies will be most telling.
|
|
