Intelligent agents via joint tabling of logic program abduction and updating

Reasoning is an important aspect for an intelligent agent to come to a rational decision. With the same importance is the ability of such an agent to adapt itself to the environment by learning new knowledge from its observations. When the agent's knowledge base is represented by a logic program, goal-directed deliberative reasoning and the adaptive ability of such an agent can be achieved by abduction and updating on logic programs, respectively. Furthermore, the tabling feature in logic programming, which affords solutions reuse rather than recomputing them, enables an agent to make an immediate decision based on past reasoning, thus avoiding repetitive deliberative reasoning. Joint tabling of logic program abduction and updating is an approach first proposed by Pereira and Saptawijaya, motivated by its application in machine ethics, enabling an agent to make moral decisions, using their system QUALM. In this paper, we provide a complete program transformation which has not been detailed on that approach. We also resolve previously unidentified issues with respect to its implementation aspects. A prototype, QUALM*, is implemented as a proof of concept using XSB Prolog. Furthermore, an application is detailed, using QUALM*, emphasizing the importance of joint tabling of logic program abduction and updating, in the context of intelligent agents, specifically in ambient intelligence for eldercare.