Learning and reasoning about interruption

By Eric Horvitz, Johnson Apacible. International Conference on Multimodal Interaction, 2003

340 citations recorded by the source corpus.

We present methods for inferring the cost of interrupting users based on multiple streams of events including information generated by interactions with computing devices, visual and acoustical analyses, and data drawn from online calendars. Following a review of prior work on techniques for deliberating about the cost of interruption associated with notifications, we introduce methods for learning models from data that can be used to compute the expected cost of interruption for a user. We describe the Interruption Workbench, a set of event-capture and modeling tools. Finally, we review experiments that characterize the accuracy of the models for predicting interruption cost and discuss research directions.

Subfields: Computer Vision and Pattern Recognition, Information Systems, Information Systems and Management

View Paper: https://www.semanticscholar.org/paper/a0a0b59106e61e0c086870693f8650bd16b1350a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Information Systems

In Information Systems and Management

Citations in the corpus

InterruptMe: designing intelligent prompting mechanisms for pervasive applications (2014, background): “While previous work in the field of interruptibility modeling identified ties between the context in which a user is, and the related level of interruptibility [18, 19, 13, 12, 37], different views on what constitutes the relevant context were not considered.”
Investigating statistical machine learning as a tool for software development (2008, neutral): “statistical models of human interruptibility [7, 12].”
Predicting human interruptibility with sensors (2005, background): “information, computer activity, and real-time analyses of audio and video streams [ 26 ].”
Interactive optimization for steering machine classification (2010, background): “Machine learning has also been used to infer such hidden user states as affect, intentions, and workload and cost of interruption from user activity, context signals, and/or physiological signals [20, 15, 16, 1216, 17].”
Psycho-physiological measures for assessing cognitive load (2010, background): “While research advances have been made in sensing context to determine when an individual is interruptible, or on monitoring the interaction between human and interrupter to understand the cost of interruption [ 19 , 20, 26], much more needs to be understood about how cognitive load factors into con”
Disruption and recovery of computing tasks: field study, analysis, and directions (2007, methodology): “Approaches explored to date include reducing the frequency and costs of interruptions through identifying the attentional state of users [13, 16, 18] and deferring or scheduling notifications in an intelligent, strategic manner [13, 20, 23], and providing support for recovering from interruption…”
Using context-aware computing to reduce the perceived burden of interruptions from mobile devices (2005, methodology): “[12] Cost of interruption based on the user’s model of attention, such as high-focus solo activity Willingness to pay to avoid the disruption”
On the need for attention-aware systems: Measuring effects of interruption on task performance, error rate, and affective state (2006, methodology): “From a hierarchical decomposition of a task, the level of a boundary could also serve as part of a non-intrusive approximation for the cost of interruption in a broader reasoning framework, such as the one developed by Horvitz and Apacible (2003). However, tuning the relationship between boundaries ”