AI Will Tell If You Are Suicidal Or Depressed
Artificial intelligence has gone a long way in helping mankind with various tasks from something mundane as opening and closing a door to extraordinary ones such as determining a person's likes and dislikes and using that data to make his or her life easier. This time, researchers have found a way to use the technology in helping humans deal with mental illness such as depression and suicidal tendencies.
As NBC News reported, suicide ranks tenth in the list of causes of death in the United States. Around 45,000 people kill themselves each year. That's 120 people a day. Those figures may be lessened with the help of artificial intelligence.
Jessica Ribeiro, a psychologist from Florida State University has been relying on technology, particularly artificial intelligence to know more about depression and suicide. Ribeiro fed the health records of 3,200 people who had attempted suicide to an algorithm which then learned the patterns leading to a suicide attempt. Her research, which is funded by the Department of Defense, has been successful so far. The researchers' methods had a 92 percent accuracy of predicting a suicide attempt within the next week and 80 percent accuracy within the next couple of years.
Another group of researchers is looking at smartphones to help in tackling depression and suicide. Cogito Corporation has come up with an app called Companion which gathers and analyzes the patterns of communication and movement from the subject's smartphone. The app doesn't listen in on the actual conversations but takes note of the calls and miscalls which are then used to develop a risk score for the individual. The clinician handling the case will know if the patient needs additional attention once the score changes.
Researchers from Carnegie Mellon University, USC, and Cincinnati Children's Hospital Medical Center went about to find a solution to this problem and hopefully prevent people from committing suicide. What they did was to study three groups of people -- those who were depressed, suicidal and a control group. The researchers took notice of all the individuals' facial expressions and gestures during interviews. The information gathered were later fed to a machine-learning algorithm which then analyzed the connection among the various expressions and gestures with regards to the three different groups.
The study came to the conclusion that smiling is the most telling gesture of a depressed and suicidal person. In the study, people who exhibited the Duchenne smile were less likely to commit suicide than those who showed the non-Duchenne smile. The Duchenne smile is highlighted by the contraction of muscles around the eyes when smiling. In other words, people whose eye muscles do not contract when smiling or when the lip corners are pulled are more likely to contemplate suicide.
Facebook has also been doing its share in addressing the issue especially after a surge of suicides and violent crimes were broadcasted on Facebook Live. the social network is utilizing different suicide prevention tools including the use of artificial intelligence to recognize patterns and determine users who are possibly depressed and suicidal.
ExtremeTech stressed that while some people claims to be good at reading another person's moods and facial expressions, they can never really tell if that person is depressed or suicidal. The signs of those emotional states are much more subtle as pointed out by ExtremeTech. This means it's quite difficult to say if someone is suffering from depression or is already contemplating suicide. Take the case of the late funnyman Robin Williams and Soundgarden frontman Chris Cornell. Their deaths took everyone by surprise simply because few or maybe even no one realized they were secretly suffering from depression. With the advancement in artificial intelligence, losses such as these would be prevented.
MORE IN ITECHPOST
Beyond Queen's Stomp-Stomp-Clap: Concerts and Computer Science Converge in New Research
The iconic "stomp-stomp-clap" of Queen's "We Will Rock You" was born out of the challenge that rock stars and professors alike know all too well: How to get large numbers of people engaged in participating during a live performance like a concert -- or a lecture -- and channel that energy for a sustained time period.
Using Waves to Move Droplets
Self-cleaning surfaces and laboratories on a chip become even more efficient if we are able to control individual droplets. University of Groningen professor Patrick Onck, together with colleagues from the Eindhoven University of Technology, has shown that this is possible by using a technique named mechanowetting.