Sometimes you don't want to be distracted by your phone. Other times it's perfectly okay. But unless you take the trouble to manually switch on and off your notifications, your device doesn't know the difference.
So engineers at Rutgers University have developed a computer model that they claim can figure out when a good time to interrupt you is. "Preferably, your smartphone would recognize your patterns of use and behavior and schedule notifications to minimize interruptions," said Janne Lindqvist.
A team led by Lindqvist, an assistant professor in the Department of Electrical and Computer Engineering , built a two stage model that predicts not only whether you're available or not, but how interruptible you are on a scale of one to five.
Common sense
Using more than 5,000 smartphone records from 22 volunteers over a period of four weeks, as well as a personality test, they were able to predict how busy people were and how they'd respond to different kinds of interruptions.
Some of the results are common sense. They found that when people were in a good mood they were more likely to be interruptible than if they were in an unpleasant mood.
They also found that people's willingness to be interrupted varies with location - at healthcare and medical facilities, people were very interruptible. But when studying or exercising, people were less open to interruptions.
More to explore
Lindqvist says that there's more to explore: "We could, for example, optimize our model to allow smartphone customization to match different preferences, such as always allowing someone to interrupt you," he said .
"This would be something an excellent human secretary would know. A call from your kids or their daycare should always pass through, no matter the situation, while some people might want to ignore their relatives, for example."
But he believes the research will eventually lead to smarter notification systems. "Ideally, smartphones would learn automatically," he said.
Smart Notifications
"As it is today, the notification management system is not smart or only depends on a user's setting, such as turning on or off certain notifications. Our model is different because it collects users' activity data and preferences. This allows the system to learn automatically like a 'human secretary,' so it enables smart prediction."
The team's peer-reviewed study, titled " How Busy Are You? Predicting the Interruptibility Intensity of Mobile Users ", will be published in May at the ACM CHI Conference on Human Factors in Computing Systems in Denver, Colorado.
0 comments:
Post a Comment