Effective coordination of distributed human planners requires timely communication of relevant information to ensure the overall coherence of activities and the compatibility of assumptions. The CODA system provides targeted information dissemination among distributed human planners as a way of improving coordination. Within CODA, each planner declares interest in different types of plan changes that could impact his or her local plan development. As individuals develop plans using a plan authoring tool, their activities are monitored; changes that match declared interests trigger automatic notification of appropriate planners. In this way, distributed planners can receive focused, real-time updates of plan changes that are relevant to their local planning efforts.

Demand and capacity imbalances in the US national airspace are resolved using traffic management initiatives designed, in current operations, with little collaboration with the airspace users. NASA and its partners have developed a new collaborative concept of operations that requires the users and airspace service provider to work together to choose initiatives that better satisfy the business needs of the users while also ensuring safety to the same standard as today. In this paper, we describe an approach to implementing this concept through a software negotiation framework underpinned by technology developed in the artificial intelligence community. We describe our exploration of peer-to-peer negotiation and how the number of conversation threads and the time sensitivity of offer acceptance led us to a centralized approach. The centralized approach uses hill climbing to evaluate airport slot allocations from a user perspective and a linear programming solver to seek solutions compatible across the user community. Our experiments with full sized problems identify the potential operational benefits as well as limitations, and where future research needs to be focused.

We describe the application of plan-recognition techniques to support human intelligence analysts in processing national security alerts. Identifying intent enables us to both prioritize and explain alert sets to analysts in a readily digestible format. Our empirical evaluation demonstrates that the approach can handle alert sets of as many as 20 elements and can readily distinguish between false and true alarms. We discuss the important opportunities for future work that will increase the cardinality of the alert sets to the level demanded by a deployable application.

We describe the application of plan-recognition techniques to support human intelligence analysts in processing national security alerts. Our approach is designed to take the noisy results of traditional data-mining tools and exploit causal knowledge about attacks to relate activities and uncover the intent underlying them. Identifying intent enables us to both prioritize and explain alert sets to analysts in a readily digestible format. Our empirical evaluation demonstrates that the approach can handle alert sets of as many as 20 elements and can readily distinguish between false and true alarms. We discuss the important opportunities for future work that will increase the cardinality of the alert sets to the level demanded by a deployable application. In particular, we outline the need to bring the analysts into the process and for heuristic improvements to the plan-recognition algorithm.