In an era when economic mineral deposits have become increasingly difficult to find, OZ Minerals and Unearthed have partnered to put the challenge to the crowd.
Making new mineral discoveries in modern mining requires innovative approaches, something OZ Minerals and Unearthed called upon for the Explorer Challenge – a global, online crowdsourcing competition.
OZ Minerals boldly released over two terabytes of private exploration data and offered up a $1 million prize pool to geologists and data scientists from around the globe, in return for ground-breaking exploration approaches that could identify targets with high prospectivity at the Mount Woods tenement near the Prominent Hill mine in South Australia.
Circulating the data on the Unearthed platform was a move that notably broke with the industry tradition of mining companies opting not to share proprietary information.
The novel approach paid off for OZ Minerals and Unearthed, with the collaboration being recognised at the 2019 Australian Mining Prospect Awards, winning the Austmine Innovative Mining Solution award.
More importantly for OZ Minerals however, the crowdsourcing competition also returned several promising targets at the copper-gold tenement that will be drilled this quarter.
OZ Minerals managing director and chief executive Andrew Cole believes the initiative has helped the company gain new insights and find fresh approaches to push the boundaries of its geological understanding of the area.
“The competition represents a fundamental change in approach to problem solving. Data science techniques can be used for exploration and many other challenges faced by the mining industry,” Cole tells Australian Mining.
“Partnering and accessing as many sectors as possible means we see ideas from outside our industry; things we may never have considered when looking with ‘mining lenses’.”
The Explorer Challenge aimed to accelerate the journey to discovery using data-driven insights by accessing thousands of skilled people from a range of technical disciplines around the world to solve a specific industry challenge.
It called on over one thousand innovators to test the limits of geology and data science by developing ground-breaking approaches to mineral exploration that can unearth new targets.
From cutting-edge machine learning to advanced physical modelling, the 37 submissions represented thousands of hours of work developing and applying robust techniques applicable to the problem of target generation.
The competition awarded first prize to ‘Team Guru’ (Michael Rodda, Jesse Ober and Glen Willis), who accepted $500,000 for their ‘interpretable machine learning models for mineral exploration using geochemistry, geophysics and surface geology’ submission.
Unearthed industry lead – crowdsourcing, Holly Bridgwater, says the winning team embraced a different way of working on the challenge.
“The winning team really took the time to deeply understand the problem and how we currently explore. This enabled them to apply a data driven approach, but one that was really relevant to the problem at hand,” Bridgwater says.
“They focussed on a solution that geologists could use. They applied machine learning and data driven approaches but made sure they were considering the geologists’ workflow. They clearly articulated the rational behind the decisions they made and how geologists can use the approach.”
OZ Minerals’ drilling program of the top consensus targets from the Explorer Challenge will include an initial six holes for 3000 metres. The mining company will follow this year’s program with further drilling in the first half of 2020.
“During this program, OZ Minerals will live stream assay data to a select group of data scientists, which will enable predictions of drilling results to be made in near real time and significantly increase the amount of information available to geologists to make faster and more informed decisions,” Cole says.
Beyond the winning ideas, OZ Minerals and Unearthed were impressed with the level of ingenuity and diversity displayed across the 37 submissions.
Both Cole and Bridgwater agree that the most surprising aspect of the competition was the wide range of machine learning techniques and workflows, as well as the use of international and national data.
Cole says the participants’ ability to explain how machine learning approaches developed relevant targets, empowering the company’s geologists to believe and engage with the method, impressed him.
“The way that data science and machine learning provide validation and feedback on predictions and models is quite different to the way we traditionally think in geological terms. We know we can learn a lot by applying this thinking to our approach,” Cole explains.
Bridgwater says all top teams across the different categories used a combination of modern machine learning techniques and more traditional approaches.
She believes mining is now more open-minded about using data through crowdsourcing initiatives following the competition and that there will be similar challenges in the future.
“There is an incentive to tap into this collective intelligence to get quick feedback on the data industry collects and move through the exploration process faster,” Bridgwater says.
“People are building trust and becoming more comfortable about data being used in a different way.”
OZ Minerals even plans to look at ways crowdsourcing can be applied to not only finding exploration targets, but also other business challenges.
“The Explorer Challenge has enabled us to push beyond the boundaries of what is normal in mining, which is how we think more broadly about innovation, and how we can apply different thinking to complex problems,” Cole concludes.