LILOG-DB: Database Support for Knowledge-Based Systems

LILOG is an interdisciplinary project for developing an experimental system performing the acquisition of knowledge by analyzing documents written in a natural language (German). Part of the experimental system is LILOG-DB, a backend database system, that supports a variety of knowledge structures as well as inferential processing. An overview of the architecture of LILOG-DB is given, focussing on the Feature Term based data model and the corresponding Feature Term Algebra (FTA). It is demonstrated how queries stated in a language based on first-order feature-logic are translated to expressions of the FTA. Finally, it is shown how LILOG-DB supports the inferential processing in the other parts of the LILOG experimental system by sophisticated rule management.