|
Artificial intelligence could be the key to finding the next source of tomorrow’s materials, according to some experts. |
mining company We are testing the potential of artificial intelligence to discover new sources of information. metal in high demandBut some experts say data limitations could hamper the technology’s usefulness.
The growing demand for metals and other materials used in electric vehicle batteries, grid-scale energy storage, power lines and other applications could hinder the world’s transition to clean energy. With this background, experts in AI and mining are looking to the rapidly advancing AI as a way to unlock these important material discoveries, especially as the use of deep learning technologies such as his ChatGPT spreads across the global economy. We are looking at technology.
Deep learning, a type of AI that includes generative AI such as ChatGPT, uses artificial neural networks to process data in a manner similar to the human brain when learning how to perform tasks and model complex systems. process. However, deep learning models rely on vast amounts of data and feedback to function, and exploration is inherently a low-feedback, intuition-based process that defines the path to successful discovery. is difficult. Additionally, the mining sector tends to be conservative in adopting new technologies.
“AI has the potential to transform the industry on many levels,” said Mining Innovation Director Joker. Axola, heavy industry technology market. “But there is great hesitation in moving forward because people are concerned about how well it works and the consequences if it fails.”
help wanted
Miners and explorers can take advantage of this assistance. T.he mining department facing A drought in the discovery of particularly important energy transition metals Copper, which is indispensable for electrification, etc. Between 1990 and 2021 he had 228 copper deposits discovered, with all but 12 copper deposits discovered before that. According to the S&P Global Commodity Insights research analysis released in May 2022, copper has made up 94.7% of metals discovered since 1990.
in the meantime, copper manufacturing Industry needs will not be met in the 2030s as consumers buy more metal-intensive electric vehicles and countries strengthen their metal-rich infrastructures. power line. that is The situation is similar for other metals, including lithium.
“With copper demand expected to outpace refined copper production, the industry is not making enough new high-quality discoveries to support a long-term pipeline,” said Sean Decoff. I’m here. Analysts at Commodity Insights said: May 2022 report.
Expanding influence of AI
Over the years, the mining industry has worked with various AI technologies, including deep learning.Miners are already applying AI and other advanced computing solutions to a wide range of projects.
For that, Freeport McMorran Companywhich AI assisted system used In 2018, it tested the factory’s capacity levels at its Baghdad mine in Arizona. The company announced in July 2020 that the project had increased copper production levels at the site by 9,000 tonnes.
Rio Tinto Group are hiring “Digital twinning” virtualization my models Test decisions on the ground before execution. The company also combined AI with analytics, machine learning and automation to improve productivity by evaluating data generated from drills, trucks, shovels, conveyors, trains and ships.
in exploration, growing rapidly JointWhat companies aim for Leverage AI with Canada’s ALS GoldSpot Discovery Ltd., Earth AI, based in Australia and the United States, Norwegian Tomra Systems ASA.
gold corpMerged with Newmont Corporation. In 2019, worked together IBM Canada develops AI software to predict where to dig in 2018 Gold Mineralization at Red Lake Mine Ontario. california-based KoBold Metals Co. you alsoses AI and Company for Mineral Exploration Recently reached unicorn status by raising $200 million in fundingThe current round is valued at over $1 billion, wall street journal Reported on June 20th.
However, recent developments in generative AI software such as ChatGPT have sparked a wave of interest in AI’s near-term capabilities, attracting the attention of investors and the industry as a whole.
AI can handle dizzying amounts of data for humans, and the continued digitization of older datasets is increasing the potential for AI-driven exploration, Kerr said. “The advantage of AI is that it can recognize what humans are simply doing. incapable I’ll take a look,” Carr said.
According to Manolis Kelis, deep-learning AI could even beat human experts in geology. Head of the MIT Computational Biology Group, of the university computer science and artificial intelligencegenerationCE laboratory.
“If you give a machine the data it needs, and the ability to learn enough representations, [and] If you can choose what expressions make sense for a particular task, you’ll beat the best geologists, or at least 90% of them,” Kelis said., looking to the near future.
wasdesert
B.hmm AI face difficult challenges in new driving discover because you’re there IThere is very little data on what leads to successful exploration, sassistance Benjamin Cox Industry consultant and mine director Open Minerals Solution. Cox said AI iswonderfulA tool that helps materialize known deposits, blocking Define and manage resources and reserves my process. Cox is one of those experts who doubts that AI will become a competent person any time soon. Prospector.
Time is also part of the problem. the discovery-to-production cycle in mines occurs over decades, averaging 15.7 years. Commodity Insights is slow to come in with new data. AI systems rely on vast amounts of information to make predictions. There is little to do compared to industries such as oil and gas, where the discovery-to-production cycle is often just a few years.. Given the relative rarity and long timelines of mineral discoveries, it can be difficult to weed out important indicators, Cox said.
“For AI to succeed, it needs to win enough to learn from it,” Cox said.
Duane Pernum, Executive chairdresser and CEO of Uranium exploration company Madison Metals Inc.He said he learned about the concept of AI in mining applications many years ago, but was nervous because of the industry’s history of protecting data. Parnum told Commodity Insig.hts or he remetersbeIns I’m skeptical about the extent to which “armchair geologists” will replace traditional mineral exploration.
“It’s much easier sInot in computer and geofantasy It’s more important to understand what’s going on, geologically speaking, than to go out there and bang rocks and really understand the geological model,” said Pernum.Using AI is great, but AI is not the only tool. “
Parnum added that only the largest mining companies may be able to do so. The computing power required to support deep learning for exploration purposes.
Feature search
Advances in deep learning’s staggering speed have constantly forced skeptics to change the goalposts that define the limits of what deep learning AI has been able to achieve over the decades.
“If I had a penny every time someone told me machines couldn’t do that, I would be a millionaire,” said Kelis.
Richard Inglis, Newmont’s chief geoscientist, has envisioned AI reimagining how exploration works to drive discovery. In the keynote speech on March 7 Canadian Association of Prospectors and Developers Annual Conference in Toronto.
Generative AI has a lot of potential that can be summarized important information Inglis said a vast array of expedition reports going back centuries has been converted into a digital database.
“The same technology can be used for mineral exploration to extract information from minerals. … Technical reports, government evaluation reports, and even the comments section of training records. ” Inglis Said, Mention Especially on ChatGPT. “And we know there’s a lot of useful information hidden there.”
in this and through others application, AI can help stem of the industry dtilted Success rate of finding more metalInglis told Commodity Insights on the sidelines of the conference.
“I hope it turns the tidesaid Mr Inglis.
S&P Global Commodity Insights creates content for distribution on S&P Capital IQ Pro.
