On the Synergy Between Plan Recognition Algorithms and Plan Based Application Systems (NOT KNOWN IF PUBLISHED BECAUSE AUTHOR HAS LEFT AT&T)

Plan recognition is an active research area in automatic reasoning, as well as a promising approach to engineering better interfaces in a wide variety of application areas. However, exactly what kind of tasks plan recognition can be used to support, and how such tasks in turn constrain plan recognition algorithms, are questions that have typically remained unexamined. In this paper we present a concrete exploration of these issues. We show how plan recognition makes interfaces more intelligent and interactive (by supporting such tasks as advice generation, task completion, context-sensitive responses, error detection and recovery), and how such tasks in turn provide representation and reasoning constraints that must be satisfied in order for the plan recognizer to efficiently (if at all) support these tasks. We then show how interfaces are fundamentally limited by current plan recognition approaches, and use these limitations to identify and motivate our current research towards a new generation of plan recognition systems. Our research is developed in the context of CHECS (CHemical Engineering Cad System), a plan-based interface system for computer aided process design.