DARPA wants to use AI to find new rare minerals

DARPA wants to use AI to find new rare minerals
DARPA wants to use AI to find new rare minerals

Secure access to rare earths is a critical national security issue because the entire United States economy is heavily dependent on minerals – and most of the minerals discovered so far are in China. The Defense Advanced Research Projects Agency (DARPA) has partnered with a company called HyperSpectral that applies artificial intelligence to spectroscopic data. This could be the key to using satellites or drones to find minerals that would otherwise be difficult to detect.

HyperSpectral CEO Matt Thereur explained how this works in an exclusive interview with Defense One. Spectroscopy is the study of the interaction of matter with light or other forms of radiation across different wavelengths. The solar radiation emitted by a particular mineral or substance due to its unique molecular composition is a unique feature.

The company has so far focused on food safety. Want to find out if large shipments of raw food contain deadly pathogens? Want to learn about the new outbreak of drug-resistant streptococci? Spectroscopy can detect bacteria that are invisible to the human eye.

“With the techniques used today, it takes a few days to tell the difference between (drug-resistant and drug-sensitive staph bacteria) because you actually have to grow the bacteria on plates and then apply antibiotics to them to see which ones they kill, if any. Today we’re talking about a swab, say from a wound, and we usually get the results within a few minutes instead of several days.”

And where does AI come into play? Thereur explains: “In nature, there are no pure samples. There is a lot of noise in nature. When we create these models with artificial intelligence, we look for all the relationships that can sometimes be obscured by the noise, for example when a section of the spectrum is disturbed by another substance in it.”

There are also multiple types of spectroscopic analysis that cannot be easily combined into a single data image. This is where AI helps, too. The auditory data derived from human speech differs greatly from text data in terms of the combinations of letters and words that are most likely to occur together. But it is their combination that makes AI-driven transcription and translation possible. In theory, spectroscopic data from a variety of sources could be just as useful.

“Whether it’s absorption, reflection, Fourier transform infrared spectroscopy, Raman or surface-enhanced Raman, it’s all about understanding the spectrographic response of these materials and being able to distinguish between different materials,” said Thereur.

What can it tell you? Thereur said the DEA used a similar technique and “was able to tell the difference between cocaine that came from one area of ​​the Colombian cartel and another.”

The cooperative agreement with DARPA is still in its very early stages, Thereur said, and the Defense Department has numerous opportunities to better understand where different materials might be located. Spectroscopy can be performed with a few specialized satellites, making it potentially useful for intelligence gathering – for example, to determine the existence of certain materials used in enemy or adversary equipment or vehicles.

The Pentagon is not only interested in gaining better access to rare earths, but also in moving the production of critical weapons and supplies much closer to the front lines, rather than relying on supply lines that would be very difficult to defend in the Pacific.

“There is a tremendous amount of applicability and use cases for spectral data analysis. Yes, there is a tremendous amount,” Thereur said.

Leave a Reply

Your email address will not be published. Required fields are marked *