After landing a contract with the U.S. Air Force in April, California-based company Labelbox announced May 20 it is making its artificial intelligence training data platform available more widely to the federal government and intelligence community.
Labelbox offers a software platform by the same name that allows development teams to manage the data used to train machine learning algorithms. For instance, in order for a machine learning tool to successfully begin identifying missile launchers in satellite imagery, it needs to be fed hundreds — or even millions — of pre-labeled pictures identifying the objects it needs to identify. The more accurate training data is fed to the algorithm, the better it works. Some government machine learning projects need to process petabytes of data per day, the company says, a flow of data that can be overwhelming.
“Labelbox is an integrated solution for data science teams to not only create the training data but also to manage it in one place,” said CEO Manu Sharma in a statement. “It’s the foundational infrastructure for customers to build their machine learning pipeline.”
The platform can be used in the cloud or on-premises, allowing data science teams to work together across agencies or locations, the company claims.
In April, the company announced it had won an Air Force Innovation Hub Network (AFWERX) Phase 1 Small Business Innovation Research contract to conduct feasibility studies on how their platform could integrate with ongoing Air Force efforts.
Now, the company is partnering with Carahsoft Technology Corp. to make their platform more widely available to the federal government and the intelligence community. Carahsoft will be able to provide the platform to interested agencies through its NASA Solutions for enterprise-wide procurement contracts and reseller partners, the company announced May 20.
Nathan Strout covers space, unmanned and intelligence systems for C4ISRNET.
