Deep learning

● Geologists have long used electron microscopes to examine patterns in grains of sand that show how they were transported (wind, waves etc.), a laborious task that often results in subjective conclusions.
● At Stanford, Mathieu Lapôtre and his team have developed SandAI, an artificial intelligence model to automate much of the analysis required to determine the geological origin of rocks, which has demonstrated 90% accuracy in testing.
● SandAI could also be used in geoarchaeology to identify the origin of materials moved by humans, and in forensic science, notably in the drive to combat illegal sand mining.