I will describe research that aims to bring human and machine closely together in an overarching model that provides a practical scaffolding for creating human-computer knowledge generation and synthesis systems. It thus addresses the fundamental issues of visual analytics: what should a comprehensive analytic reasoning model contain, how should those parts be linked together, and what is the role of interactive visual interfaces in controlling the model? This research provides a roadmap for generating relevant knowledge and appropriate decisions from data of such size and complexity that key entities, relationships, and contexts could not otherwise be uncovered, understood, and acted upon.