Extending an assumption based truth maintenance system to databases

Abstract

A description is given of how information stored in databases can be used to support reasoning using an assumption-based reasoning system with truth maintenance. The basic notion behind this type of reasoning is that the problem solver can draw inferences on the basis of assumptions in the absence of information to the contrary. These inferences are nonmonotonic, in the sense they may be contradicted by update to the information base and hence lose their validity status. The objective is to describe some techniques for maintenance in databases to extend an assumption based truth maintenance system to make use of large amounts of data stored in databases. There are two levels of truth maintenance taking place in the system. Truth maintenance occurs at the level of the database and at the level of explicit facts derived or asserted. The focus is on the computational techniques that are necessary at the former level.<>