Context Aware Support System and Intuitive Interfaces (CASSINI)

For intelligence analysts, rigorous and timely analysis of structured and unstructured data demands efficient collection, integration, exploration, and exploitation of multi-dimensional data, across a range of missions and contexts. Unfortunately, the ability to support these tasks is severely limited by current tools that rely extensively upon costly manual processing by analysts and inefficient workflows for interacting with automated tools. To address these needs and develop improved human-on-the-loop workflows, we propose to design and demonstrate Context Aware Support System and Intuitive Interfaces (CASSINI). CASSINI will combine advanced multi-dimensional information visualizations with deeply integrated collaboration tools and novel reinforcement learning automation to support exploration and reasoning in analyst workflows. First, we will extend upon prior cognitive analyses conducted with intelligence community practitioners to define the requirements for joint analyst-automation teaming. Next, we will design a set of ecological human machine interfaces and collaboration tools within a fusion-oriented streaming analytics workflow environment. Finally, we will extend existing visualization environments to rapidly prototype CASSINI interfaces and workflows. Combined, CASSINI-based support tools will facilitate intuitive visualization, interaction, and teaming with new automated models as part of effective and efficient analyst workflows.