The Pentagon transfers the Maven project, a renowned artificial intelligence initiative, to the NGA
DENVER — The National Geospatial-Intelligence Agency is evaluating the progress of “Project Maven” as it prepares to take over the artificial intelligence initiative and integrate it into a broader range of efforts to apply the machine learning to geospatial intelligence.
The Biden administration is proposing to move the Maven project to NGA as part of its fiscal year 2023 budget request. The program has been managed by the Office of the Secretary of Defense since its inception in 2017.
NGA Director Vice Admiral Robert Sharp said the agency will “engage the industry” regarding the transition of the Maven project in the coming months. The agency has repeatedly stressed in recent years that it will need to use artificial intelligence and machine learning to process and analyze the growing amount of satellite imagery and other GEOINT data available from government and commercial sources.
“We want to move forward together, so that we can deliver GEOINT at the pace that our warfighters and decision makers need,” Sharp said during a keynote Monday at the GEOINT conference here. “We need to be able to keep up with rapidly changing digital trends. We must be able to accelerate our ability to deliver detections at mission speed, to give our customers a tactical, operational and strategic advantage.
NGA has partnered with Project Maven since its inception, helping to provide images and other data needed by companies to train their algorithms, according to Mark Munsell, NGA’s deputy director of data and digital innovation.
And NGA has also worked on computer vision and machine learning projects, according to Munsell. He said NGA can “benefit from everything Project Maven has learned over the past five years and integrate software projects into its own infrastructure.”
“It makes a lot of sense to bring these things together,” Munsell said during a Tuesday media roundtable with reporters on the fringes of GEOINT. “That doesn’t mean we’ll always operate the same way Maven operated. We will make a very good assessment of what they have done. We call on our experts in the field, who have been impregnated with GEOINT for 30 years, some of them. They are now the ultimate customer to be able to assess utility, assess how we can continue to support our combat support partners in the military, and make changes to improve it.
The agency also hopes to avoid duplicate efforts, while sharing promising software between the military and intelligence components it supports.
“If you have developed an algorithm that succeeds after certain objects in certain geographies and biomes, we will leverage that as a community and ensure that if someone else in the community needs it, we can provide as a service to them,” Munsell said.
Military AI Scout
Project Maven was created in 2017 “to accelerate DOD’s integration of big data and machine learning,” Undersecretary of Defense Bob Work wrote in a memo at the time. The project fielded its first algorithms for processing images and video captured by surveillance aircraft later that year.
The program served as an AI pioneer for the Pentagon, predating the creation of the Joint AI Center.
It also made headlines in 2018 when thousands of Google engineers protested the company’s involvement in the project. The company ended its involvement, later saying it would not develop AI applications in the fields of weaponry or surveillance.
But Project Maven has continued to grow well beyond the controversy. The Pentagon requested $247 million for Cross-Functional Algorithmic Warfare Teams, aka Project Maven, in FY2022 after receiving $230 million for the program in FY21. It was also among programs to drive the use of “colorless” software appropriation. .
It’s unclear how much the Pentagon is asking for the program in FY23 as it transitions to the NGA, as details of the intelligence community’s budget are classified.
But the Department of Defense budget documents provide details on the progress of Project Maven in recent years. The program aims to “augment and automate” the processing, exploitation and delivery of full-motion video feeds from a range of unmanned aerial vehicles, including “WAMI ISR”, which stands for Intelligence, Surveillance and Reconnaissance by large-scale imaging.
The program also uses AI to automate the analysis of military and commercial satellite imagery, according to the documents.
The program has also expanded beyond imaging in recent years. It also uses AI to mine “CEM”, which stands for “capturing enemy material” in intelligence parlance, as well as “maritime” and “PAI” intelligence, or publicly available information.
“Maven’s AI, deep learning and computer vision algorithms and ideas are developed for use in theater to detect, classify and track objects in images (e.g. people , vehicles, and weapons) as well as to provide other information, such as with CEM, text-based and other projects,” the documents state.
NGA is taking over the project’s “GEOINT AI services and capabilities,” according to Sharp, and it’s unclear what will happen to the capabilities the project has created for other categories of intelligence.
With the transition not expected to go into effect until Oct. 1, the NGA is currently doing the “administrative work” to figure out how to transition the contracts, as well as getting staff and management in place, according to Munsell.
And even as the NGA assesses how Project Maven fits into its broader portfolio, Munsell also stressed that there will be “no pause” in the project’s ongoing activities.
“Assessment is not a break,” he said. “It is our responsibility as a functional GEOINT manager to help prioritize, to help others understand investments, and then with data and statistics, to provide opportunities for improvement.”