Interruption Management in Web-based Collaborative Systems

Interruption management has received a good deal of attention from researchers, but this work has not had a substantial impact on real-world applications. This paper examines the factors associated with interruption management in web-based systems (a highly restricted class of environments) and presents an adaptive interruption management system, Interrupt-Broker, as a component of a web-based visual analytics toolkit. Interrupt-Broker seeks to find the right time for delivering messages in such a way as to minimize the cost of interruptions. These costs include distraction from the primary task and disruption of cognitive processes through premature incorporation of new information in the course of sensemaking (e.g. conformity to a collaborator) in a collaborative analysis and decisionmaking system. A hybrid deliberative/ reactive architecture, inspired by enactive cognition theories, is designed to deal with minimal sensory information gathered from the user interaction to determine users' cognitive state. This architecture implements reinforcement learning and reactive action selection.