Turn Data into Insights

Across the federal government, agencies and their mission partners are using innovative approaches to solve complex and emerging problems – and they’re doing it with data resources they long had in-hand.

The difference is that today, advances in technologies like machine learning, artificial intelligence, and high-performance computing are dramatically changing what agencies can do with their data.

Leveraging data to make better, smarter or faster decisions has the potential to transform the way government operates. These new tools at agencies’ disposal are as powerful, accessible, and varied as ever before, and with them, new missions as well as current missions in be supported in entirely new ways.

Machine Learning to Determine Safe Drone Flight Paths

Earlier this year, a commercial pilot flying near Los Angeles International Airport (LAX) spotted a drone within 700 feet of the plane—dangerously close. Fortunately, the plane avoided the drone without incident. Unfortunately, these incidents are becoming far too common.

There are currently nearly 900,000 drones registered with the Federal Aviation Administration (FAA) and more than a quarter million remote pilots have been certified to fly them. This creates an increasingly growing concern of the proliferation of unmanned aircraft systems (UASs) near U.S. airports.

A solution to address this growing problem turns the immense amount of data collected in domestic airspace into insights for drone safe flight areas. Working with the FAA, federal systems integrator General Dynamics Information Technology (GDIT) developed a UAS Risk Analysis Model that determines safe flight zones and routes for drones. The model uses altitude and GPS coordinates to predict flight paths and other data in areas not covered by radar coverage, often the lowest altitudes near airports where planes are taking off and landing.

GDIT applied machine learning to massive amounts of geospatial data in building the model. Billions of flight path data points at 850 U.S. airports were analyzed and predictive analytics is used to simulate activity in areas not the areas not covered by radar.

There are 20 airports now evaluating this model, paving the way for a nationwide implementation at the 30 busiest airports in the United States.

AI to Advance Health Care and Access

Skin cancer is the most common cancer in the U.S. and can be even more common for our nation’s service members and veterans. This makes early screening and treatment even more critical to improve health outcomes.

Working together with the Department of Veterans Affairs (VA), GDIT built an AI-based classifier to help spot potential malignant skin lesions from photos. Deep learning AI image classification models were trained on more than 10,000 images of skin lesions to classify images of skin lesions into seven common categories, determine if it’s indicative of a common skin disease, and recommend follow-up care.

Once field tested, veterans and providers will be able to use the app as part of telehealth services to review photos of skin lesions. Physicians will be able to use the tool to more rapidly and accurately diagnose skin conditions, which may have gone undetected during delayed or deferred care during the COVID-19 pandemic.

Detecting Fraud, Waste, and Abuse in Medical Claims

And at the Centers for Medicare and Medicaid Services (CMS), machine learning is being used to identify potentially fraudulent Medicare payments. Each day, CMS receives millions of claims representing billions of dollars in payments for healthcare services. Using machine learning, outlier detection and behavioral analysis, the agency is able to identify risky providers and, in turn, analyze large volumes of data that previously went unanalyzed.

This is a huge endeavor, with similarly huge impact, detecting and preventing hundreds of millions of dollars in healthcare fraud and combatting other forms of waste and abuse.

“Deriving value from data can transform the way agencies deliver services,” GDIT Vice President, AI and Data Insights Dave Vennergrund said. “Ultimately, this positively impacts the missions they advance and the people they serve.”