Distributed belief revision as applied within a descriptive model of jury deliberations

摘要: Belief revision is a well-research topic within AI. We argue that the new model of distributed belief as discussed here suitable for general modelling judicial decision making, along with extant approach known from jury research. The to interest, whenever attitudes information are be simulated multi-agent environment agents holding local beliefs yet by interaction with, and influencing, other who deliberating collectively. In proposed, it's entire group agents, not an external supervisor, integrate different opinions. This achieved through election mechanism, principle "priority incoming information" AI models problematic, when applied factfinding jury. present incorporates computable revision, such recoverability adopted. By this principle, any previously held must belong current cognitive state if consistent it. For purposes simulation calls refinement. Yet we claim, it constitutes valid basis open system where functionalities (or outer stiumuli) could attempt handle aspects deliberation which more specifi legal narrative, argumentation in court, then debate among jurors.