network/cyber
The Pentagon is committed to growing artificial intelligence. (Graphic by Breaking Defense, original brain graphic by Getty)
Washington – The National Geospatial Intelligence Agency (NGA) wants to build a platform to assess supply chain risks posed by artificial intelligence and machine learning technologies. Project Maven is an initiative aimed at accelerating the use of AI across the military.
According to the new NGA mandate, the platform must map all vendor links to each other and discover “troubled relationships within the vendor network.”
“Maven Needs [a] It is a rapid, open-source, unclassified reporting system that allows NGA to share information of interest to vendors and governments without inadvertently disclosing vendor-proprietary information,” the April 24 invitation read. says. “The platform can also serve as a repository for mission reports and information sharing on advanced and critical threats within the agency’s supply chain.”
Project Maven will become an official program of record in fiscal year 2024. The request requires NGA to assess the risks of Project Maven’s AI supply chain. critical components and personnel.
“Furthermore, unraveling the AI/ML supply chain will allow us to assess the health of AI/ML vendors and identify the opportunities, vulnerabilities, and systemic dependencies that are critical to making sound and secure prototype and acquisition decisions. We can identify it,” the solicitation said.
NGA has listed specific tasks that a supply chain risk management platform can perform. This includes rapid network mapping of vendor relationships, organizing data labeling efforts, documenting unclassified information shared with vendors, and “algorithm-based technology and program integration.” of Record is a rapid prototyping effort that adapts, prototypes, and integrates commercial AI technology into her DoD platform with active tactical users. ”
Founded in 2017, Project Maven is deviated Last year, he moved from the Office of the Undersecretary of Defense for Intelligence and Security to the NGA. As the Department of Defense explained at its creation, the project “focuses on computer vision (an aspect of machine learning and deep learning) to autonomously extract objects of interest from moving and still images. . […] Biologically inspired neural networks are used in this process, and deep learning is defined as the application of such neural networks to learning tasks. ”
The NGA conducts operational management of the program’s GEOINT services, but also the Pentagon’s premier digital and AI office. exclusive Project Maven is part of President Joe Biden’s 2023 budget request to Congress.
Last April, then-director of the NGA, Robert Sharp, said a key area of focus for the agency was expanding Project Maven to combine “computer vision” with human expert analysis. rice field.
“Starting next fiscal year, there will be complementary computer vision efforts that will provide automated GEOINT detection to both intelligence analysts and fighters,” Sharp said at the time. “As a key role as a combat support agency, we provide subject matter experts – humans who can train machines to evaluate machines and understand their output. Bringing together diverse and sometimes siled communities of data scientists, GIS professionals, and imagery analysts to improve AI model performance, develop standards, and lead interoperability efforts in the GEOINT community. To do.”
