Representing and Aggregating Conflicting Beliefs

Title: Representing and Aggregating Conflicting Beliefs

Research Question: How can we represent and aggregate conflicting beliefs in a multi-agent system, especially when dealing with sources of varying reliability?

Methodology: The researchers propose a modular, transitive relation for collective beliefs. This allows for the representation of conflicting opinions and has a clear semantics. They compare their approach with quasi-transitive relations often used in Social Choice. They also describe a way to construct an agent's belief state by aggregating information from informant sources, taking into account their reliability. This construction circumvents Arrow's Impossibility Theorem, which is a significant contribution to the field. Finally, they present a simple set-theory-based operator for combining the information of multiple agents, ensuring that the operator is well-behaved when iterated and computationally effective.

Results: The researchers propose a representation for collective beliefs that distinguishes between group indifference and group conflict. They also provide a mechanism for an agent to construct its belief state by aggregating information from sources of varying reliability, which circumvents Arrow's Impossibility Theorem. Additionally, they present a well-behaved operator for combining the information of multiple agents.

Implications: The research has important implications for multi-agent systems. The proposed representation and aggregation method provide a more accurate and flexible way to deal with conflicting beliefs and varying source reliability. The well-behaved operator for combining agent information allows for more efficient and robust decision-making processes in these systems. Overall, the research contributes to a better understanding of how to represent and manage belief states in multi-agent systems.

Link to Article: https://arxiv.org/abs/0203013v1 Authors: arXiv ID: 0203013v1