The fitness tracker that knows if you're faking: Researchers train smartphones to spot cheaters - Rehabilitation Institute of Chicago

Skip to Content

RIC in the News

Published on January 12, 2016

Source: DailyMail

The fitness tracker that knows if you're faking: Researchers train smartphones to spot cheaters

By Cheyenne Macdonald 

You may be able to lie to your doctor about how much you exercise each week, but you can't lie to your fitness tracker—at least, not anymore.
Researchers in Chicago have come up with a way to train these easily deceived devices to spot fake activity.
This could allow healthcare providers and insurance companies to get a more accurate scope of a patient's exercise habits.
Shaking your phone while binge-watching TV may once have passed as a brisk walk for standard smartphone fitness trackers, but the new system can differentiate between what's real exercise, and what isn't. This could allow healthcare providers and insurance companies to get a more accurate scope of a patient's exercise
Shaking your phone while binge-watching TV may once have passed as a brisk walk for standard smartphone fitness trackers, but the new system can differentiate between what's real exercise, and what isn't. This could allow healthcare providers and insurance companies to get a more accurate scope of a patient's exercise
HOW THEY TRAINED THE PHONES
In the study, 14 subjects between the ages of 23 and 38 used various strategies to trick their devices.
The participants shook the phone while sitting to pretend they were walking.
To make the motions seem more realistic, they swung their hands, or slipped the phone into their pockets and moved their body around.
They also attempted to trick the smartphone in the opposite way, pretending to sit while they were actually walking.
The systems were found to be better at detecting the cheaters: true activity was predicted with 38 percent accuracy, but the ones trained on deceptive behaviour increased accuracy to 84 percent.
Researchers at Northwestern Medicine and Rehabilitation Institute of Chicago (RIC) trained smartphones on normal and deceptive activity.
Shaking your phone while binge-watching TV may once have passed as a brisk walk for standard smartphone fitness trackers, but the new system can differentiate between what's real exercise, and what isn't.
The systems were found to be better at detecting the cheaters: true activity was predicted with 38 percent accuracy, but the ones trained on deceptive behaviour increased accuracy to 84 percent.
While the study focused on smartphone trackers, it could also be applied to wearable fitness devices.
'As health care providers and insurance companies rely more on activity trackers, there is an imminent need to make these systems smarter against deceptive behaviour,' said lead study author Sohrad Saeb, a postdoctoral fellow at the Center for Behavioral Intervention Technologies at Northwestern University Feinberg School of Medicine.
'We've shown how to train systems to make sure data is authentic.'

The researchers explain that the more accurate system can help health care provides monitor their patients, to see if they are following advice to become more active.

And, some insurance companies will offer discounts to active individuals.

In the study, 14 subjects between the ages of 23 and 38 used various strategies to trick their devices. The participants shook the phone while sitting to pretend they were walking.

To make the motions seem more realistic, they swung their hands, or slipped the phone into their pockets and moved their body around.

The researchers explain that the more accurate system can help health care provides monitor their patients, to see if they are following advice to become more active. And, some insurance companies will offer discounts to active individuals

The researchers explain that the more accurate system can help health care provides monitor their patients, to see if they are following advice to become more active. And, some insurance companies will offer discounts to active individuals

They also attempted to trick the smartphone in the opposite way, pretending to sit while they were actually walking.
'Very few studies have tried to make activity tracking recognition robust against cheating,' said senior author Konrad Kording, a research scientist at RIC and an associate professor in physical medicine and rehabilitation at Feinberg.
'This technology could have broad implications for companies that make activity trackers and insurance companies alike as they seek to more reliably record movement.'
The system was trained to recognize the behaviour of one person, and generalize it to work on many individuals.
The trackers were retrained up to six times during the trials.
Still, Saed says the system isn't perfect quite yet.
'If someone attaches an activity tracker to a dog, the system can't recognize that.'


For full article, including illustrations, video, and comments, go to: http://www.dailymail.co.uk/sciencetech/article-3396194/The-fitness-tracker-knows-faking-Researchers-train-smartphones-spot-cheaters.html#ixzz3x9Pa4cwf