A network for closed-world default reasoning

To account for lack of knowledge, closed world default (CWD) reasoning has being an important method in human reasoning to infer knowledge when its converse does not exist. The paper proposes a hybrid network with CWD rules aiming at providing a better and more realistic framework for default reasoning. In this system, either positive CWD rules or negative CWD rules can be selectively or globally applied to the knowledge. For that knowledge without CWD rules, their reasoning mode will remain unchanged and they co-exist with those with CWD rules in the same framework.