The paper is about trust in information sources in the context of Multi Agents Systems and it is focused on information and trust propagation. Trust definition is inspired from Cognitive Science and it is seen as a truster's belief in some trustee's properties which are called: sincerity, competence, vigilance, cooperativity, validity and completeness. These definitions are formalized in Modal Logic and it is shown that even if trust, in that sense, is not transitive, we can find interesting sufficient conditions based on trust that guarantee that the truth of an information is propagated along a chain of information sources.
展开▼