Goldcorp has shared updates on how IBM’s Watson technology has improved their mining operations over the past year.
In early March 2017, IBM Canada announced it would be working with Goldcorp to use Watson’s technology in the Canadian mining industry for the first time as a mineral explorer. The majority of this work would take place at Goldcorp’s Red Lake, Ontario mine. Early results have shown that IBM’s Watson Explorer helped reduce the time for processing data from 165 hours to 4.5 hours.
“The ability of IBM Watson to understand, learn, reason, and interact will have a considerable impact on the Canadian mining industry by enabling geologists to make more data-driven decisions, faster,” said Mark Fawcett, a partner with IBM Global Business Services, back in 2017. “In industries such as natural resources where knowledge professionals are working with a rapidly growing volume of data, Watson is helping to evolve how professionals operate, make decisions, drive value, and gain competitive advantage.”
Thus far, Watson has been used to analyze vast amounts of data, from land survey information to drilling reports. This has helped scientists and miners find specific areas to explore next, reach high-value exploration targets quicker, calculate geological models with more certainty, and interpret the growing amounts of data that comes with new discoveries.
Goldcorp had a lot of data to go through—years and years of drill logs along with 140 Block Models with tons of paper files. The team was spending up to 80 per cent of their time looking for and manipulating data, when they should be using that time to analyze instead.
Phase one of the Watson integration was completed in lat 2017, and it involved educating the system, setting up the foundation, and making sure the integrity of the data was verified.
“Structured data from a wide range of sources was inputted, including drill logs, geological models, and block models, together with unstructured analysis and maps, to provide insights into Red Lake’s geology, current and historic mining and exploration activities, and successful exploration techniques,” reads a release.
Watson was able to institute a 97 per cent efficiency gain when it was tasked to help with the above data. This progress has allowed Goldcorp geologists to find new information from existing data and hone in on results, greatly assisting the exploration process by making the areas of focus smaller.
Now that phase one is over, phase two is well underway. The purpose of this is for Watson to identify new data correlations, come up with new conclusions in other datasets, improve on predictions from areas without a lot of data, and find better zones of mineralization.
If Watson can improve on these aspects and complete a successful phase two, it will result in higher drill success rates at existing mines, meaning less new mines started and less of an environmental impact.
Goldcorp is a gold producer with a large portfolio of mines. They wanted to combine Watson’s natural language processing and machine learning aspects to help them analyze a trove of legacy data and free up geologists to do their job better.