Gian Piero Zarri , CNRS-National Research Council, France
Elisa Bertino , UniversitÓ degli Studi di Milano, Italy
Session: 6.2-6.3 Intelligent Database Systems
This tutorial provides a "state of the art" about the integration of Artificial Intelligence (AI) and Database (DB) techniques. We will first outline some insufficiencies of the "traditional" (relational) DB technology when dealing with new applications such as Computer Aided Design and Manufacturing (CAD/CAM), office applications, Computer Aided Software Engineering (CASE), military command and control, etc. We will then classify the studies concerning AI/DB integration into two categories, i.e., a) efforts originated in a prevailing DB context; b) efforts originated in a prevailing AI context. Accordingly, we will describe, in the first part of the tutorial, the nested relations systems, the semantic and hyper-semantic data models, the semantic data modelling approaches such as OMT (Object Modelling Technique) and UML (Unified Modelling Language), and the Object-Oriented Data Base Systems (OODBMSs); these last will be explained in particular detail. In the second part, we will introduce first the main knowledge representation systems used in the AI domain: the resolution principle, logic programming, rule programming, inference by inheritance, frames and description logics, etc. These knowledge representation systems are the AI equivalent of the data models used in the DB domain. We will speak then of deductive databases, of integration of DBMSs and Expert Systems shells, and of "advanced solutions" like TELOS, Conceptual Graphs, NKRL and, mainly, the well-known (and very controversial) CYC system by D. Lenat. For each type of existing, "integrated" system, we will detail some case studies and we will discuss the commercial realisations. In the final part of the tutorial, we will introduce a general overview of some very 'hot' topics that are influencing deeply both the AI and the DB field. We will then speak here of temporal databases, ontologies, structured, semi-structured and unstructured data, mediators and wrappers, multi-agent systems, data mining etc.