source:GREEK REP RTER
A groundbreaking machine-learning model has been developed to predict where minerals can be found on Earth, and even on other planets. This technological breakthrough holds tremendous significance for both the scientific community and various industries.
Scientists and researchers are constantly seeking to uncover the secrets of our planet’s history and extract valuable resources like those used in rechargeable batteries. By analyzing patterns in mineral associations, this innovative model has the potential to revolutionize mineral exploration and enhance our understanding of celestial bodies.
Shaunna Morrison and Anirudh Prabhu led a team with the goal of creating a technique to identify the presence of specific minerals. This objective has typically been viewed as more of an artistic skill than a scientific one. In the past, it has relied heavily on individual expertise and a fair bit of good fortune.
What did the team achieve?
The team successfully developed a machine learning model that utilizes data from the Mineral Evolution Database. This comprehensive database contains information on 295,583 mineral locations and covers 5,478 distinct mineral species.
By analyzing association rules within this data, the model can predict the presence of minerals in previously unexplored areas. This breakthrough opens up new possibilities for uncovering unknown mineral occurrences.
Validation of machine learning model to find rare elements
To validate the model, the researchers conducted tests in the Tecopa basin, an area in the Mojave Desert known for its resemblance to Mars.
Remarkably, the model successfully predicted the presence of several geologically significant minerals in this region. These included uraninite alteration, rutherfordine, andersonite, schröckingerite, bayleyite, and zippeite.
This achievement demonstrates the model’s ability to accurately identify important minerals in real-world environments, showcasing its potential for advancing our understanding of both Earth and other planetary bodies.
Identification of areas with high probability of rare earth elements
The AI model successfully pinpointed areas with high potential for critical rare earth elements and lithium minerals. Notably, it identified promising locations for minerals such as monazite-(Ce), allanite-(Ce), and spodumene.
This capability of analyzing mineral associations holds immense value for professionals in the fields of mineralogy, petrology, economic geology, and planetary science.
The authors highlight that mineral association analysis can serve as a powerful predictive tool, enabling researchers to make informed decisions and advancements in their respective domains.
Scope of the method
The scope of this method extends beyond just mineral associations. It can be applied to analyze the coexistence of fossils, microbes, molecules, and other characteristics within geological environments.
The versatility and applicability of this association analysis approach make it highly valuable and influential in various fields of data-driven research, focusing on the evolving Earth and planetary systems.
Additionally, there is an exciting opportunity to explore the combination of mineral occurrences with microbial data, incorporating their physical, chemical, biological, and geological parameters.