How a university researcher is using machine learning to help identify suicide risk
Philip Resnik was a computer science undergrad at Harvard when he accompanied a friend to her linguistics class. Through that course, he discovered a fascination with language. Given his background, he naturally approached the topic from a computational perspective.
Now a professor at the University of Maryland in the Department of Linguistics and the Institute for Advanced Computer Studies, Resnik has been doing research in computational linguistics for more than 30 years. One of his goals is to use technology to make progress on social problems. Influenced by his wife, clinical psychologist Rebecca Resnik, he became especially interested in applying computational models to identify linguistic signals related to mental health.
“Language is a crucial window into people's mental state,” Resnik said.
With the support of Amazon’s Machine Learning Research Award (MLRA), he and his colleagues are currently applying machine learning techniques to social media data in an attempt to make predictions about important aspects of mental health, including the risk of suicide.
Developing more sophisticated tools to prevent suicide is a pressing issue in the United States. Suicide was the second leading cause of death among people between the ages of 10 and 34 in 2018, according to the Centers for Disease Control and Prevention (CDC). Among all ages that year, more than 48,000 Americans died by suicide. Resnik noted the COVID-19 pandemic has further increased the urgency of this problem via an “echo pandemic.” That term has been used by some in the mental health community to characterize the long-term mental health effects of sustained isolation, anxiety, and disruption of normal life.
