Collaborating with Multiple Distributed Perspectives and Memories

Experience-based reasoning is a powerful tool for decision-support systems which attempt to leverage human memories as examples to solve problems. However, there is also a requirement to leverage recorded experiences that are created by a set of diverse users from a variety of work positions, with varied goals, cultural backgrounds, and perspectives across time. New computing methods need to be specified and defined in order to resolve some of the problems introduced by these requirements. The research discussed in this paper focuses on the ongoing issue of how to reconcile the different personal and cultural perspectives of agents in order to make group decisions. We discuss how individual agent history and culture can influence experience-based reasoning. We compare our approaches to traditional experiencebased reasoning techniques in order to highlight important issues and future directions in social decision making. Results from an experiment are presented that describe how shared problem solving presents issues of consensus and buy-in for various stakeholders and how a constraint based coherence mechanism can be used to ensure coherence in joint action among a group of agents and improve performance.