Part of the reason the company focused its initial efforts on Canada is that the country has large amounts of survey data in the public domain, including narrative field reports, ancient geological maps, data. geochemicals on drill hole samples, airborne magnetic and electromagnetic survey data. , lidar readings and satellite imagery spanning decades of exploration.
“We have a system where we can ingest all of this data and store it in standard formats, control the quality of all the data, make it searchable and be able to access it programmatically,” Goldman explains.
The high-tech momentum
Once they’ve compiled all the information available for a site, the KoBold team explores the data using machine learning. The company could, for example, create a model to predict which parts of ore deposits have the highest concentrations of cobalt, or create a new geological map of an area showing all of the different rock types and fault structures. It can add new data to these models as it is collected, allowing KoBold to adaptively modify its “near real-time” exploration strategy, Goldman explains.
KoBold has already used knowledge from machine learning models to acquire its Canadian mining claims and develop its programs in the field. Its partnership with Stanford’s Center for Earth Resources Forecasting, underway since February, adds an additional layer of analysis to the mix in the form of an AI “decision agent” that can plot a full exploration plan.
Stanford geoscientist Jef Caers, who oversees the collaboration, explains that this digital decision maker quantifies the uncertainty in the KoBold model results, then designs a data collection plan to sequentially reduce that uncertainty. Like a chess player trying to win a game in as few moves as possible, the AI will aim to help KoBold make a decision on a prospect with minimal wasted effort, whether that decision is to drill in a spot. particular or to move away.