Interactive Knowledge-Goal Reasoning

Knowledge goals are used by reasoning entities to fill in information required for decision making and are an important part of computational understanding systems. Simple knowledge goals can be solved through traditional information retrieval techniques or database queries. In order to solve complex knowledge goals, however, a plan must be composed and executed in an investigative manner. During computational investigation, goals of both the system and the user can change, resulting in goal trajectories that may inform future goal solutions. In this paper, we propose an interactive reasoning system that leverages a case-based methodology to solve complex knowledge goals represented as natural language queries.