DARPA Seeks AI Solutions for Sourcing Critical Minerals

DARPA has initiated a competition aimed at discovering AI solutions for the sourcing of critical minerals. These minerals are crucial raw materials necessary for producing items essential to national security.

In collaboration with the US Geological Survey (USGS), DARPA is investigating how machine learning and AI technologies can expedite the assessments of critical minerals. Dr. Anne Fischer, Deputy Director of DARPA’s Defense Sciences Office, stated, “The USGS’s critical mineral resource assessments are fundamental to our home supply and production of these crucial materials. Our goal is to significantly enhance the USGS’s capabilities, particularly in areas vital to national security.”

Currently, the evaluation of the existing list of 50 critical minerals is a labor-intensive process that does not meet the urgency of modern supply chain demands. The AI competition encompasses two main challenges:

• Map Georeferencing Challenge: Participants will access a dataset comprising 1,000 or more various maps to train and validate models, aiming to link coordinate points to known locations on base maps.
• Map Feature Extraction Challenge: Participants will receive a training set with maps, where each legend item is labeled and categorized (as point, line, or polygon), along with a binary pixel map showing the feature’s coverage. The challenge is to identify all features in a map that appear in its legend.

Sarah Ryker, associate director for energy and mineral resources at the USGS, remarked, “DARPA is renowned for cultivating innovation by forming research communities focused on solving specific problems, and we are eager to apply the results of this competition in real-world scenarios. The US is currently under-mapped, and with the Bipartisan Infrastructure Law, we have a rare chance to enhance our mapping efforts effectively.” Each challenge will offer prizes: $10,000 for first place, $3,000 for second, and $1,000 for third.

While the financial incentives may not be substantial within the AI industry, the opportunity to contribute to impactful projects is a motivating factor for participants. Registration for the map georeferencing challenge is now open, while the map feature extraction challenge registration begins on August 29. Additional information and challenge registration can be found on the official website.