How AI technology promises to aid in finding gold
By Robin Fell | 2 minute read | November 15, 2018
In my job as the Director of Strategic Technology Solutions at Goldcorp, a leading Canadian mining company, I am keenly aware that our industry lags behind most others in terms of the use of technology. Sadly, the joke is that in mining, innovation means bigger trucks.
At Goldcorp, we’re on a journey of innovation. We’re willing to invest in new ideas and do things differently, in a way that perhaps our industry has never done. For example, we really care about finding technological solutions that will help us improve safety and protect the environment. We know that to innovate we need to work with companies that are as good at technology as we are at mining.
Exploiting technology to improve mining operations
Most of the ‘easy to find’ gold deposits have already been found and mined; discovering new sources of ore is a very slow, challenging and expensive process. We need more innovative ways of finding new economic gold mineralization. As an industry, we must use our assets more efficiently to improve our success.
Traditionally, mining companies have thought of their assets in terms of people, machinery and institutional knowledge. However, we have billions of dollars invested in data assets that we have not even begun to tap. Right now, it’s very hard to interpret all the data we’ve collected to enable us to improve our decision making.
We believe that the key to unlocking the value in this data lies in harnessing the capabilities of advanced analytics systems, including AI and deep learning, that can turn big data into useful insights.
Mining for data insights
We were already working with IBM in other areas of our business, so the concept of using IBM Watson and AI technology to extract more value from our data intrigued us. Now we’re using Watson’s machine learning and deep learning capabilities to predict gold mineralization, focusing on Goldorp’s Red Lake Gold mines.
First, we’re providing Watson with our structured data from diverse sources such as 2D/3D and drill data. This insight can be linked to unstructured data hidden away on localized servers, allowing us to accelerate geological analysis. And we’ve developed predictive models using deep learning that enables us to make predictions on gold mineralization to be interpreted by geologists.
Extracting value from big data
Today, value creation in mining has shifted from how well we can move material to how well and quickly we can act on information to increase efficiency and production. Watson can ingest and analyze the data in ways and at speeds we could never accomplish. Our geologists can find patterns and insights from our data that previously remained hidden away like the deepest of ore deposits.
Even when we have the capability, Watson can answer questions faster – averaging 35 times faster. And then we can ask follow-on questions that Watson can answer based on information it already has, and the answers it has already given. So, we’re essentially increasing the power of insight exponentially, so our geologists can make the most informed decisions in the least amount of time.
Watson isn’t reinventing mining or replacing the institutional knowledge of our best geologists. It’s augmenting the ability of the geologist, which ultimately benefits our operations overall.
IBM has the technology and people to drive innovation and Goldcorp has the knowledge and experience in geology and mining. Together, we’ve built something that will benefit and transform our industry.