Mineral Exploration: Geologists Versus Artificial Intelligence

Geology, once bound by traditional methods, is experiencing a seismic shift thanks to the integration of artificial intelligence (AI) and machine learning (ML). These cutting-edge technologies are unlocking a wealth of possibilities in the field, empowering geologists to analyze vast amounts of data, make accurate predictions and optimize decision-making processes. 

“ML enables geologists to see a new vision, new sets of data and come up with new ideas that we had not thought of before,” says John-Mark Staude, CEO, Riverside Resources.

Several crucial properties must be considered during the analysis of geochemical data, including statistical distributions that deviate from normality, heteroscedasticity (unequal variances indicating different distribution modes), spatial non-stationarity models and the compositional nature of data. Accurately accounting for these factors is vital to ensure reliable interpretations and meaningful results in geochemical studies and exploration projects. AI and ML algorithms play a significant role in handling these complexities and extracting valuable insights from diverse geochemical datasets.

Geology focuses a wire range of highly diverse data, from a continental scale to small, individual rocks. The challenge is to go from a larger scale to a specific scale to find a feasible target, that is one of the key areas where MI and AI can help. “Visualization software allows us to see huge amounts of data and filter it more rapidly, which in the end results in spending fewer days in detailing and more days in actually visiting potential deposits,” says Staude.

ML harnesses the power of geosciences through the recognition and utilization of complex patterns in high-dimensional data, enabling valuable inferences. Another significant advantage of ML is its capability to handle data variability, including that arising from utilizing multiple data sources. Through appropriate algorithm selection and careful control of model overfitting, machine learning can effectively mitigate the effects of data variability and deliver reliable results. 

Despite the advantages of these technologies, it is necessary to be careful with data inputs, as supervising and ensuring high-quality databases is of great importance to ensure accurate analysis. However, bada data is better than no data. “It is important to note that low-quality data and obtaining biased results can also help companies in better training their algorithms,” says Roman Teslyuk, Founder and CEO and CTO, Earth AI.

The main struggle when developing AI and ML applications is the lack of data. To address this issue, governments play a crucial role in generating and providing updated information. Geological surveys play a significant role in enabling databases to be analyzed. Those datasets are of great value to mining companies of all sizes, and would benefit mining projects in Mexico. “Mexico should support and build databases to feed all the algorithms, understanding how different geological terrains respond. These things can drive these sciences to the future,” says Don Hulse, Director of Business Development, Forte Dynamics. 

The integration of AI and ML does not replace the expertise and insights provided by geologists, however. Instead, these tools complement human knowledge, enhancing exploration and analysis capabilities. Geologists have valuable domain expertise, geological intuition and contextual understanding that AI and ML algorithms lack. The collaboration between geologists and these intelligent technologies leads to more robust interpretations, improved decision-making processes and the ability to explore and analyze vast amounts of data efficiently. 

“Finding a new ore deposit is like finding a needle in a haystack, every method could bring us closer to that goal. But we also need to be mindful that we can not just let something run loose without geologists on the ground. I do not think that AI will replace geologists but it is beneficial for the main experts in geology to partner with the main experts in other areas to advance our knowledge,” says Katharina Pfaff, Associate Research Professor, Colorado School of Mines.