Interruption Timing Prediction via Prosodic Task Boundary Model for Human-Machine Teaming

Exploring the Effects of Interruptions in Different Phases of Complex Decision-Making Tasks

“…The phase information can be either labeled by the human teammate proactively (e.g., clicking interface buttons representing different phases) or inferred by the information system through analyzing coarse-grained interaction activities, such as the interfaces the human teammate is visiting in our experiment. Compared with the difficulty to identify subtask boundaries in a hierarchical task model, which requires continuously monitoring of user or interface events and demands high computation power (Katidioti, Borst, Bierens de Haan, et al, 2016; Peters, 2020; Peters et al, 2019), the coarse-grained division of phases can provide considerable performance protection at a lower implementation cost. Furthermore, these two approaches are compatible and can be used together if conditions permit the application of subtask-boundary-identification methods.…”

“…The former approach requires specialized equipment and can hardly be applied out of laboratory environments. The latter approach entails a detailed task analysis and relies on observable cues such as computer operations (e.g., mouse clicks, opening and closing windows), overt movement or expression of users (e.g., head movement, spoken communication), and visual and acoustical analyses of the physical task environment (Peters, 2020;Peters et al, 2019). This approach is mainly useful for tasks consisting of a clear sequence of subtasks that ities, and the developed models are limited to this approach in complex decision-making contexts is limited because the task involves many activities without overt observable cues, such as reading a research report from the screen.…”