Reasoning about Truthfulness of Agents Using Answer Set Programming

This paper describes a declarative framework for representing and reasoning about truthfulness of agents using Answer Set Programming. The paper illustrates how, starting from observations, knowledge about the actions of the agents, and the normal behavior of agents, one can evaluate the statements made by agents against a set of observations over time. The paper presents an ASP program for computing the truthfulness of statements of agents over time. It also introduces the notions of a supporter (denial) of a statement and discusses possible ways for computing minimal supporter (denial). The paper presents an application of the proposed framework, specifically in detecting man-in-the-middle attacks targeting computer and cyber-physical systems. Finally, the paper briefly relates the proposed framework to works on trust and reputation of agents and discusses possible extensions.