RallyPoint, a professional social media network serving the U.S. military and its veterans, today announced a pilot program in collaboration with Amazon Web Services from Amazon.com Inc. and researchers at Harvard University’s Nock Lab to help veterans with issues of self-harm and suicide.

According to the U.S. Department of Veterans Affairs, an estimated 17 military veterans die by suicide each day. This is a rate 50% higher than seen among nonmilitary adults in the general population.

As a major source of camaraderie and social interaction for military members, RallyPoint felt that the company needed to help address this problem. That’s especially so because its platform often becomes a convergence point for veterans and military families seeking companionship and help from people in similar situations.

“Suicide is a national public health emergency facing our nation,” said Dr. Barbara Van Dahlen, executive director of the President’s Roadmap to End a National Tragedy of Suicide task force, PREVENTS for short. “Sadly, our veterans are at high risk — often because of the sacrifices they have made to protect our freedoms. Because suicide is not just a concern for our veterans, it’s critical that leading veteran and nonveteran organizations partner in order to ensure that everyone is protected against this very real threat.”

RallyPoint connects nearly 2 million service members, veterans and their families with a space where they can feel free to discuss military and veteran life, share information, swap stories and find resources geared toward military members. The company also offers numerous helpful resources, including critical mental health and suicide prevention information.

Although it makes up an extremely small fraction of the discussions that occur on RallyPoint, some members bring up depression, thoughts of self-harm and suicide. As the platform of a broader community, the company wanted to make it a priority to help these individuals by providing them timely assistance and information that could give them a lifeline.

In order to speed the discovery and correct identification of individuals in need of these resources, RallyPoint and the Amazon Machine Learning Solutions Lab turned to a machine-learning model using Amazon SageMaker and Amazon SageMaker Ground Truth, an AI system that can categorize and label incoming data based on training criteria. The company trained the system using anonymized public posts and then sent the model along to experts at Harvard’s Nock Lab, a group of researchers who study human suicidal behavior, to improve the accuracy of the model.

The model uses a human-in-the-loop system to better tune its results by surfacing potential insights and presents them to actual people in order to incorporate their judgment. For example, as the ML platform works, it will discover posts potentially revealing information about self-harm and a person reviews it to tell the difference between apparent humor and someone who is in need of help.

A spokesperson at RallyPoint told SiliconANGLE in an email that before this integration, the platform used several manual methods for discovery including a robust flagging system, such as the option for members to flag public posts or bring posts to administrators’ attention. A strong administrator system and dedicated community moderators who are active on the site worked with direct messages brought to light by other people.

“The user will see a message with support information posted to the public discussion where the indication of potential self-harm is exhibited, and which could include support group or hotline information,” the RallyPoint spokesperson said. “The individual will also receive a private message with similar information from a trained RallyPoint employee. The purpose of this outreach is to adhere to the protocols that RallyPoint put in place under counsel from the Department of Veterans Affairs and researchers from Harvard to increase the chances that members will take steps to get help.”

This new AWS-driven machine-learning driven resource will do a very similar thing and hopefully catch patterns missed by human participants, giving moderators and administrators a chance to intervene in situations that could have gone unattended in the past.

“As the technology provider of the platform facilitating these interactions, we believe it is our responsibility to make it easier for our members to navigate the sea of goodwill when they publicly share that they have a problem they seek to solve,” said Dave Gowel, a military veteran and chief executive of RallyPoint.

With this new mechanism available on RallyPoint and the platform’s already broad measure of mental health resources – including the Wellness Hub – the company hopes to continue to show a dedication to reducing suicides within the military community.

To date, the pilot has proved successful. The learning model now quickly surfaces potentially sensitive posts to the RallyPoint and Harvard teams and reduces the overall need for time-consuming manual review to quickly deploy life-saving intervention and information.

“We have been studying suicide for a hundred years, but unfortunately, suicide rates remain high,” said Matthew Nock, chair of psychology at Harvard University. “One reason for this is that most people who are feeling suicidal don’t seek out traditional mental health care. We believe that the use of technology can help revolutionize the prevention of suicide, allowing us to identify and provide resources to people in real-time when they’re experiencing distress.”

Photo: Pixabay