|
Main
Description
Status
Research
Publications
Demo
People
Funding
Links
|
Research
KANAL checks process models specified by a user reports possible errors in
the models, and generates specific suggestions to the user about how to fix
those errors. Our approach is inspired on previous work on EXPECT using Interdependency Models (Kim and Gil
AAAI-99, Kim and Gil
AAAI-2000). KANAL analyzes the interdependencies among the individual
steps of a process model. For example, a simulation of the execution of the
steps allows KANAL to anlayze interdependencies between the conditions and
effects of different steps, such as that the required conditions for each step
are met when the step is supposed to take place, and that the expected effects
of the overall process are in fact obtained.
The kinds of checks performed and the kinds of fixes proposed by KANAL include
- Unachieved Preconditions
- Unachieved Expected Effects
- Unordered Steps
- Inappropriate Execution of Steps
- Inappropriate Loops
- Inappropriate Disjunctive Branches
- Inappropriate Branches
The KANAL has been used by
various end users including biologists who built and tested models of complex
molecular biology processes and army officers who used KANAL to critique their
plans, i.e., courses of action (COAs). We found that its analysis report was
very useful for checking the plans that were entered by users and often
suggested useful ways of improving the plans. However we also found that
sometimes users wish to extend the background knowledge that is used to
produce the analysis report. For example, they may want to define various
special cases of actions that are relevant for different situations, or
exceptions that were not specifically addressed in the general background
knowledge. In general, knowledge bases are never complete and it is important
to provide a capability of adapting the knowledge base for varying needs. To
support this capability we have extended our authoring environment so that the
user can choose to enter special cases of critiquing knowledge as well as to
change the description of the process itself in order to improve the resulting
report. In this extension, instead of describing different cases in a single
definition, we make use of the inheritance mechanism in the given knowledge
representation system to represent each special case as a separate entity. We
believe that this approach provides a more a natural view of special cases
for end users and has efficiency advantages. The details are described in
Kim and
blythe, IUI-2003.
|
