Blog: Mining Firm Expands Its Reach with Artificial Intelligence – Jutia Group
Artificial intelligence (AI) is used in many sectors and applications, but not so much in the mining space. That makes no sense because AI is an extremely useful tool in dealing with challenges that have tremendous amounts of data and involve hundreds of variables. However, AI is just a tool, it’s not a silver bullet solution. For mining, the area most in need of what AI has to offer is exploration for new discoveries.
The costs of exploration are going up. The easy stuff has already been found, the higher-grade ore already mined. Most exploration programs are searching in part, or entirely, for mineralized zones that are “undercover” (under overburden). Overburden is worthless rock or soil overlying a mineral deposit. It’s the enemy of exploration geologists because it’s very difficult to “see through.” Various tools are deployed to understand what’s beneath the overburden: geophysics; geochemistry, etc.
But each property is unique. What works well in one place might not work at all in another. Exploration (not just for metals) is the perfect setting—high uncertainty, lots of data, lots of variables, the potential for a big reward and substantial cost savings—for AI to be a valuable, cost-effective tool to aid geologists. Albert Mining’s (AIIM:TSX.V) technology will never replace geologists, only assist them. Income tax software has been around for >20 years, but it has not replaced tax accountants—not even close.
Albert Mining has been using AI, machine learning and data mining for less than a decade. Clients benefit from a multidisciplinary team that includes professionals in geophysics, geology, AI and mathematics. Most of the team has been together for six to 10 years; they have 30+ proven discoveries. Management says it has a 70% success rate in identifying new zones of mineralization. If a gambler could be right just 60% or 65% of the time, he or she would become quite wealthy. Albert Mining’s success rate is not higher because sometimes there’s no additional mineralization to be found. Or, there’s not enough data for the technology to operate at an optimal level, or the geology is simply too difficult to understand.
Each time the company’s CARDS (Computer-Aided Resources Detection System) finds something, it saves the client considerable exploration time and money, and frees up managerial resources. The benefits are many, for all stakeholders, when there’s a successful outcome.
Let me throw in this testimonial by Ron Perry, then a director of Metanor Resources, that I paraphrased from a recently shot video [see 2-minute video clip]. This was an Albert Mining success story from 2009-10 that’s going into production. How many exploration projects make it to production? Not many!
Ron Perry, formerly of Metanor Resources: “I met Michel [Fontaine] of Albert Mining in 2009, He came into our booth at the Cambridge show, it sounded like he had some proprietary algorithms. We hit it off. I approached our VP of Exploration. I told him, ‘If we use Albert Mining’s technology and find something, it’s your discovery. If not, just blame me, the finance guy’. . .In 2010 we made a discovery, off the main road, in an area that had no inkling or smell of gold. At the same time, Michel’s team did some mapping around our tailings pond and there was another discovery, and that discovery is now going into production. A pretty amazing technology. . .”
What’s Albert Mining’s Secret Weapon?
Albert Mining owns 100% of a proprietary software package called CARDS, a state-of-the-art computer system used to identify areas with a high statistical probability of containing mineral deposits. High statistical probability is a key phrase: CARDS does not provide certainty; no system of any kind can. The backbone of CARDS is a knowledge extraction data mining engine that uses pattern recognition algorithms to learn the signatures of positive and negative data points to create a model that makes predictions on the positive or negative nature of new data points. CARDS uses powerful algorithms to analyze digitally compiled exploration data and identify zones with a high potential for discovery.
How Does All of This Work?
Data is entered into CARDS in the form of a geo-referenced database. Each data point, or “cell,” in the database is linked to its own set of criteria extracted from geophysical surveys, drill and rock sample assays, geological maps, etc. It’s critical that as much data as possible is captured and logged. Importantly, the system can take many different forms of data, (geophysics, geochemistry, topography, satellite, geology, faults, spatial data, etc.) which is why it’s such a robust predictive tool. The data is divided into two databases.
The first database includes cells with known assay results (drill hole/rock samples data, etc.) and is used to develop—to learn—a model of the geological target being sought. The model could be set to identify targets containing >5 g/t gold. The second database, equally important, includes cells with no assay results. Complex algorithms are used to identify those cells that have a high similarity to the signatures of positive mineral deposits. In addition, in the analysis of each cell, the characteristics of cells within a specified distance, in the same neighborhood, are weighed into the evaluation of that cell.
Therefore, even cells lacking data can be effectively evaluated if the combination of their limited characteristics, and their proximity to cells with other significant characteristics, is similar to that of cells with known positive results. Unlike rule-based computer models, CARDS is not biased by the rules of any particular geologic model. In fact, because of CARDS’ ability to learn and make predictions based on the signatures of multiple positive data points, it can make predictions on any geological deposit type.
Once the data has been crunched, which typically takes several weeks, prospective targets generated by CARDS are evaluated with the client, and both parties discuss and outline potential exploration and drill targets.
New Business Model, New Chairman, New Sectors: A New Albert Mining
In the past, this is where Albert Mining’s work ended. They would provide a valuable service and move on. However, all of that is about to change. The company, with the support of large shareholders, is taking ownership stakes in companies it does work for. In addition, a new chairman should be named shortly. He or she will be well versed in this new operating model. The company’s objective is to develop new revenue streams by participating in the exploration success of its AI-assisted exploration.
Albert Mining is currently running a program in Norway for Playfair Mining (PLY:TSX.V). Management recently invested CA$100,000 in Playfair at CA$0.05 (2 million shares) and signed a CA$75,000 service agreement. The RKV project covers two past-producing volcanogenic massive sulfide (VMS) copper mines (Kvikne and Rostvangen), a magmatic nickel-copper deposit (Vakkerlien) and >20 additional mineral occurrences. Management will use its proprietary technology to analyze a large amount of geophysical, geochemical and geological data to uncover patterns hidden in Playfair’s data. It will then run those machine-learned patterns through its algorithms to identify prospective targets.
This technology has been successful in assisting geologists in the identification of a number of major mineral discoveries, especially in the context of VMS mining districts. As management continually tweak CARDS, they will now benefit more directly, and more substantially, from their high success rate.
CEO Michel Fontaine tells me his company will look a lot different in three or four months. A new chairman, a new business model, perhaps programs outside of just mining. Readers may recall that in a recent interview with Michel he mentioned searching for abandoned underground land mines and exploring for water as possible new business segments. Any high-value field or application where there’s a lot of complex data is ideally suited for CARDS.
As CEO Michel Fontaine said in our interview last month: “We feel this is just the beginning of something potentially much bigger. No other company has the breadth and& depth, the vast experience in AI technology for mining (and new sectors) that we do. No one.”
Peter Epstein is the founder of Epstein Research. His background is in company and financial analysis. He holds an MBA degree in financial analysis from New York University’s Stern School of Business.
The content of this article is for information only. Readers fully understand and agree that nothing contained herein, written by Peter Epstein of Epstein Research [ER], (together, [ER]) about Albert Mining, including but not limited to, commentary, opinions, views, assumptions, reported facts, calculations, etc. is not to be considered implicit or explicit investment advice. Nothing contained herein is a recommendation or solicitation to buy or sell any security. [ER] is not responsible under any circumstances for investment actions taken by the reader. [ER] has never been, and is not currently, a registered or licensed financial advisor or broker/dealer, investment advisor, stockbroker, trader, money manager, compliance or legal officer, and does not perform market making activities. [ER] is not directly employed by any company, group, organization, party or person. The shares of Albert Mining are highly speculative, not suitable for all investors. Readers understand and agree that investments in small cap stocks can result in a 100% loss of invested funds. It is assumed and agreed upon by readers that they will consult with their own licensed or registered financial advisors before making any investment decisions.
At the time this article was posted, Peter Epstein owned no shares of Albert Mining and the Company was an advertiser on [ER]. Readers understand and agree that they must conduct their own due diligence above and beyond reading this article. While the author believes he‘s diligent in screening out companies that, for any reasons whatsoever, are unattractive investment opportunities, he cannot guarantee that his efforts will (or have been) successful. [ER] is not responsible for any perceived, or actual, errors including, but not limited to, commentary, opinions, views, assumptions, reported facts & financial calculations, or for the completeness of this article or future content. [ER] is not expected or required to subsequently follow or cover events & news, or write about any particular company or topic. [ER] is not an expert in any company, industry sector or investment topic.
Streetwise Reports Disclosure:
1) Peter Epstein’s disclosures are listed above.
2) The following companies mentioned in the article are billboard sponsors of Streetwise Reports: None. Click here for important disclosures about sponsor fees. The information provided above is for informational purposes only and is not a recommendation to buy or sell any security.
3) Statements and opinions expressed are the opinions of the author and not of Streetwise Reports or its officers. The author is wholly responsible for the validity of the statements. The author was not paid by Streetwise Reports for this article. Streetwise Reports was not paid by the author to publish or syndicate this article. Streetwise Reports requires contributing authors to disclose any shareholdings in, or economic relationships with, companies that they write about. Streetwise Reports relies upon the authors to accurately provide this information and Streetwise Reports has no means of verifying its accuracy.
5) From time to time, Streetwise Reports LLC and its directors, officers, employees or members of their families, as well as persons interviewed for articles and interviews on the site, may have a long or short position in securities mentioned. Directors, officers, employees or members of their immediate families are prohibited from making purchases and/or sales of those securities in the open market or otherwise from the time of the interview or the decision to write an article until three business days after the publication of the interview or article. The foregoing prohibition does not apply to articles that in substance only restate previously published company releases.
Graphics provided by the author.