Advancements in machine learning applications for mineral prospecting and geophysical inversion: A review

Mineral deposits are a globally important resource. However, the supply of shallow minerals is close to depletion, forcing exploration activities to expand to deeper areas. Deep exploration has greater challenges compared to shallow exploration, and how to effectively extract the intrinsic connection between exploration data and deep concealment, and how to quickly and accurately locate target zones, remain urgent challenges to be solved. Mineral prediction and geophysical inversion are the core links in mineral exploration, and how to make up for the shortcomings of traditional methods in these links has become an important topic of current research. In the past decade, with the wide application of big data technology in the field of geological prospecting, more and more geological data have provided support for the application of machine learning (ML) in geophysical exploration and mineral prediction. ML overcomes the limitations of traditional methods to a certain extent, such as reducing human subjectivity and improving the ability to mine the laws among geological data, showing great potential. This study summarizes the progress of the application of ML, especially deep learning, in the field of mineral exploration in recent years, focuses on the two key aspects of geophysical inversion and mineral prediction, analyzes the advantages and limitations of the various methods, and makes concluding comments on the future direction of development, with the aim of providing valuable references for the on-site application of ML in mineral exploration and the direction of future research.