From Kolflow Project

Revision as of 07:20, 7 February 2011 by Desmontils-e (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Title : Inferences and Interactions for dealing with Semantic Inconsistencies

Code : 3

Responsible : Orpailleur

Activities : Design and experiment strategies for dealing with Semantic Inconsistencies

Start Date :

End Date :

Objectives : Given a knowledge base KB expressed in a classical logic, such as OWL DL, it is important that KB is semantically consistent: if it is not, anything (and its negation) can be produced by classical deductive inferences. Now, given two (or more) sources of knowledge, such as a human and a knowledge discovery process, leading to two consistent knowledge bases KB1 and KB2, their conjunction KB1 ∪ KB2 may be inconsistent. If this occurs, what should be done? It is the objective of this task to give some answers to this question, when the knowledge bases are ontologies expressed in a description logic-like formalism.

Success criteria : The ultimate goal of this task would be to demonstrate a system piloted by a user that is able to manage inconsistencies: ideally, the role of the user should only be to make choices among proposals of the system.

Risks : The theoretical complexity of revision operators satisfying all the AGM postulates is proven to be high, even for propositional logic. Moreover, there is, to our knowledge, no known algorithm for revision in expressive description logic that satisfy all these postulates. However, there exist algorithms that should be easy to adapt in order to do an approximate revision, satisfying only a part of the postulates. It remains to be shown that they behave adequately for the purpose of Kolflow.

Deliverables :

description Feb2011+months
D31 algorithms and architecture of the interactive semantic inconsistency managing system 1212
D32 prototype and tests of the interactive semantic inconsistency managing system 2424
D33 prototype of the extension integrating management of ”human-machine” and ”logical” inconsistencies 3636

Sub-tasks :

Task31 Revision and Merging of Ontologies.
Task32 Human-Computer Interactions for Managing Consistency.
Task33 Interactive Revision and Merging of Ontologies.
Task34 Interactive Reasoning with a Globally Inconsistent Family of Ontologies.

Participants :