Advancing new discoveries: creating better data for ML in geoscience

Geoscience machine learning

Novacene has recently partnered with the Geological Survey of Newfoundland and Labrador (GSNL) to evaluate its data readiness for current – and future – machine learning

Machine learning (ML) is increasingly being used across many different industries – from health care to business, from finance to retail.

It has also become an important tool among geoscientists. In recent years, ML was used to successfully predict the presence of gold mineralization in Manitoba.

This surprising discovery is one example of how ML opens a world of possibilities for the field. This revolutionary technology can reveal new information in untouched areas that are difficult to access – leading to potentially groundbreaking geological discoveries. But there’s one thing that currently stands in geoscientists’ way: data that is compatible with ML technology.

Since early 2022, the Geological Survey of Newfoundland and Labrador (GSNL) has been determined to break down this barrier for geoscientists – and partnered with NovaceneAI’s experts to ensure clients can effectively use their public data with ML to unlock Earth’s incredible secrets.

The “data dilemma”

GSNL keeps all of its data in the Atlas – a primary resource that geoscientists use for several aspects of their work, including mapping areas and creating accurate models to help with their exploration.

While it’s a precious tool that geologists rely on to conduct their work, the data hasn’t quite kept up with ML technology. Therefore, ML technology has a difficult time ingesting this important information.

“At Novacene, we believe this is an exciting time for geoscientists as they begin to incorporate ML into their important work. But first and foremost, they need reliable data that can work alongside ML,” says Marcelo Bursztein, founder and CEO of NovaceneAI.

Becoming the user to find solutions

GSNL handed over their Atlas to the NovaceneAI team, who was tasked with taking this data and advising on how to make it compatible with ML technology. While it might sound like an overwhelming job, Novacene knew how to tackle it: they partnered with a professional geoscience firm to examine the data through the eyes of the actual users.

“We approached the project from the viewpoint of a company that wanted to explore new mine sites in unexplored or underdeveloped areas, and required GSNL public geoscience data,” Bursztein explains.

They specifically examined the types of data that a geoscientist would need to conduct their work. As they went through the data, the Novacene team determined whether the data was ready to work with AI, whether a machine learning technology could read it, and the Atlas’s overall data accessibility.

Helping geoscientists overcome data roadblocks

When Novacene finished its review, the team developed a list of helpful recommendations for GSNL. This advice will help guide the GSNL as it works to ensure that its data is compatible with both current, and future, ML.

The Novacene team also discovered some useful tips for geoscientists themselves – creating an assessment they can use when they need this data, and apply it to their ML tools.

“We wanted our purpose to be two-fold: helping the GSNL bring their data up to date so it can be compatible with ML, and developing a resource that GSNL’s clients can use themselves. We want geoscientists, and the GSNL, to be able to work together so they can both reach their full potential.”

Unlocking the potential in Newfoundland and Labrador

Newfoundland and Labrador has so much untapped potential in the world of geology. According to the GSNL, it is a “highly competitive jurisdiction for mineral resource development” – even though it has poor bedrock exposure, a relatively short field season, and a lack of access in remote areas.

ML technology can finally break down these barriers, and allow geologists to uncover all that Newfoundland and Labrador has to offer. And, potentially help them make a groundbreaking discovery.

“Our work with the GSNL will help them remain ahead of the curve. Novacene’s recommendations will also provide the basis for the creation of a new, modernized Atlas – a precious resource that geoscientists can rely on for decades to come,” Bursztein says.