Mining and Resources

Even slight efficiency improvements in mining have significant implications for a company's bottom line. When applied in a targeted manner, machine learning and artificial intelligence has been proven to push the boundaries of resource detection and extraction logistics.

  • Going Deep for Mineral Treasure
    Javen Shi

    Professor Javen Shi, Director of Advanced Reasoning and Learning

    Prominent Hill, in South Australia’s far north, has been a rich source of minerals for almost 20 years. Keen to extend the life of this resource, mine operator OZ Minerals created a crowd- sourcing competition to pinpoint potential exploration sites.

    The OZ Minerals Explorer Challenge took place over three months, involving more than 1000 participants from 62 countries.

    They dug through more than six terabytes of public and private exploration data to identify mineral deposits and find new ways to access them.

    There was a $1 million prize pool on offer as well as the prestige of leading the way for the local mining industry.

    Professor Javen Shi led a team from AIML and the University’s Institute of Mineral and Energy Resources on the treasure hunt, in collaboration with industry experts in minerals exploration and geoscientific modelling.

    His DeepSightX consortium exploited multi-disciplinary skills at the intersection of artificial intelligence and geoscience to analyse the exploration data sets.

    AIML provided machine learning techniques and engineering support, while the Geoscience team members contributed an understanding of the exploration process, industry best practice and true domain expertise.

    “The team developed a drilling exploration plan that took advantage of the overwhelming data available, while being justifiable from a geoscientific perspective,” Javen says.

    “We achieved this by integrating the latest concepts from mineral systems modelling, with recent breakthroughs in deep learning – artificial neural networks and algorithms inspired by the human brain that learn from large amounts of data – and computer vision.”

    The result was a world-class predictive modelling capability, which confidently recommended a series of drilling targets and enabled feedback of more data to further develop the model and AI targeting.

    The DeepSightX team took out second place in the OZ Minerals Explorer Challenge, winning a $200,000 prize.

    The results of the international challenge offer the potential to revitalise mineral exploration and discovery in South Australia.

    Javen’s team was proud to further the cause of integrating AI into the mining industry and plans to expand and commercialise its work.

    “The competition was a prelude for DeepSightX and we look forward to the exciting journey ahead,” Javen says.

    The DeepSightX team has embraced the opportunity to develop deep learning to improve mineral exploration and help uncover the treasure in our own backyard.