Domain-independent reason-enhanced controller for task-oriented systems-DIRECTOR
AAAI'04 Proceedings of the 19th national conference on Artifical intelligence
| Hi-index | 0.00 |
Agents need to operate in dynamic situations where the information they have about the world is incomplete, uncertain and quite possibly false. Active logic has been designed with capabilities that enable these aspects of the world to be taken into account, notably time-situatedness, contradiction tolerance and meta-reasoning. This paper presents a general-purpose implementation of active logic, Alma/Carne that is meant to be a basis on which to build and experiment with such agents. We illustrate the use of Alma/Carne in the implementation of a non-monotonic reasoner that is computable and that has been successfully tested on a large number of examples from the literature.