Metal Tech News - The Elements of Innovation Discovered

Mining technology critical green energy electric vehicles rare earth metals minerals news

By A.J. Roan
For Metal Tech News 

Time to give AI mineral exploration a try

Minerva provides unique solution amidst COVID-19 pandemic Metal Tech News Weekly Edition – April 29, 2020

 
Series: COVID-19 | Story 11

Last updated 6/27/2020 at 5:52am

Cognitive AI machine learning for mineral exploration Minerva Intelligence

Adobe Stock

AI-assisted mineral exploration provides exploration geologists unable to get into the field due to COVID-19 a means of generating new exploration targets with previously collected geological, geochemical and geophysical information.

With many mineral exploration geologists self-isolating behind a computer screen instead of out in the field breaking rocks, artificial intelligence offers a means of generating new exploration targets with the reams of previously collected geological, geochemical and geophysical information.

Minerva Intelligence, which is redefining the emerging industry of artificial intelligence-assisted mineral exploration, believes the TERRA Mining AI suite of software services it has to offer can position exploration companies for greater success once COVID-19 restrictions are lifted.

"In light of our current COVID-19 reality, I feel that now is the best time to address database quality issues that companies have potentially been putting off for years," Sam Cantor, head of the Economic Geology Department at Minerva Intelligence told Metal Tech News. "Now is the time to get your house in order and utilize the best available tools for desktop exploration, as most of us can't be in the field."

Minerva has developed four interrelated AI-driven services to help organize, predict and generate exploration targets prior to returning to the field.

LEO – a document management system for organizing and searching for specific data. Using AI software, LEO uses controlled vocabularies to automatically index geo-locations and tag files, making document management and searching more efficient.

DRIVER – a specialized software system that finds critical multi-element zones in drilling results for exploration and geometallurgy by evaluating multi-element geochemistry against existing mineral deposit knowledge.

TARGET – a cognitive AI system that produces explainable exploration targets by mimicking the traditional process of geological data evaluation. Input data is aligned with geological rules and is compared to dozens of expert-based mineral deposit models.

SOLACE – consultation and software service that unites data from multiple projects under a single, auditable standard. As project data is often arrayed in databases that are not interoperable, SOLACE restructures vocabularies to fix typos, alternative labels and legacy codes.

Each of these products in Minerva's TERRA Mining AI suite can be utilized on its own or in tandem to provide the most comprehensive and effective approach to any future stake.

Science of interpretation

One of the most common AI techniques used for processing big data is machine learning, a self-adaptive program that gets increasingly better at analysis and pattern recognition with experience or with newly added data.

Most AI systems today, however, are based on deep learning, where adaptation happens through exposing the AI system to tens of thousands of illustrative examples.

This method involves absorbing intricate details and subtle nuances in pictures, videos, or sounds into the parameters of the neural network of the AI system.

Deep learning AI works well for operating mine tasks such as autonomous haul truck driving, process optimization and predictive maintenance for equipment.

The reason deep learning is successful for these types of applications is because there is a lot of continuous data, as well as control over that data.

Most of this information tends to be more objective, hard numbers – RPMs of a vehicle, ambient temperature or when exactly a haul truck has broken down. The most important aspect of this data is very clear indicators of success and failure.

Geology, a science of interpreting data generated by Earth's processes millions of years earlier, however, requires the ability to translate subjective data and correlate it to the available objective data.

Things such as the geophysical information, geochemistry, rock types, mineral systems, and location in the world need to be factored in and this can be troublesome as there is no uniform standard to help geologists quickly utilize the data.

A certain level of assumption, or an inference of data closer to human problem-solving abilities, is necessary.

This is the type of cognitive AI that Minerva has developed for mineral exploration.

Minerva's cognitive AI

The rise of AI is likely to be one of the most significant trends in the technology sector in the coming years as companies of all sizes and in various sectors look to improve decision-making, reduce operating costs and enhance consumer experience.

One of the leading voices in the AI world, Gary Marcus, is calling for more technology and expertise in the AI field that Minerva operates in, cognitive AI.

"Recent research in artificial intelligence and machine learning has largely emphasized general purpose learning and ever larger training sets and more and more compute," Marcus penned in a paper on AI over the next decade. "In contrast, I propose a hybrid, knowledge driven, reasoning-based approach, centered around cognitive models, that could provide the substrate for a richer, more robust AI than is currently possible."

Recent breakthroughs in AI have been achieved by applying machine learning to very large datasets.

However, machine learning – in particular, deep learning systems – often fails when there is limited training data available or when the actual dataset differs from the training set.

Also, it is often difficult to get clear explanations of the results produced by deep learning systems.

Minerva's AI platform addresses these limitations by using human-generated knowledge models and probabilistic reasoning to generate predictive analyses.

This is accomplished by using semantic networks to store information for their computer reasoning.

Semantic networks represent logical relations between concepts in data and allow cognitive AI to closer replicate how a human would think about subjects.

Using semantic networks as its primary data format means Minerva's software is specifically engineered to represent conceptual knowledge and not just unique traits in a dataset.

This allows their system to also provide explanations in natural language for predictions as well as expert advice about what exploration measures should be taken.

A unique trait – and something that gives most geologists a headache when trying to organize and extrapolate data for any given exploration project ¬– is a lack uniformity or standardization in the nomenclature and data previously collected.

According to Mining World Magazine, geologists can spend up to 80% of their time searching for and preparing data and the remaining 20% interpreting and analyzing it.

This means a lot of time, effort, and especially money goes into preparing the required information for exploration, and before even drilling can occur companies want to be as close to absolutely sure they are making a sound investment.

An individual has a working knowledge model of how things are similar: this is an innate quality of the human mind. Therefore, by better mimicking how exactly a person derives information, a kind of synergy is made from this new AI direction.

This is the basis for Minerva's cognitive AI.

Desktop exploration

Due to the COVID-19 pandemic, field programs for many companies have been significantly – and sometime entirely ¬– limited.

As a result, field geologists who would rather be investigating outcrop on a promising exploration target have the luxury of time to carry out "desktop exploration," which could be taken a step farther with the assistance of AI.

"Smart allocation of time now can mean that when the boots return to the ground at some point, you are going to be chasing well-thought-out, robust exploration targets," Cantor said.

With a vast accumulation of geological data and the expertise of years of scientific insight, the company's TERRA Mining AI is essentially creating a virtual assistant that can be added to a mineral explorer's toolbox.

Shawn Ryan, who is renowned for his discoveries across the White Gold District in Canada's Yukon Territory, has incorporated Minerva's TERRA AI into his exploration efforts.

Economic Geology Department at Minerva Intelligence AI mineral exploration

Sam Cantor

"By using Minerva's cognitive AI to identify the geochemical relationships between various structures, we can essentially get an unbiased second opinion to augment our own exploration efforts – and at a price that's about the cost of a single RAB hole," said the founder and chief technical officer of White Gold Corp.

With more prospectors and geologists spending more time exploring from their offices than in the field, now may be the best time for the mineral exploration sector to investigate the potential of AI.

EDITOR'S NOTE: More information on the use of AI in Yukon's White Gold District can be read at White Gold adds AI to exploration system in the April 15 edition of Metal Tech News.

 

Reader Comments
(0)

 
 

Our Family of Publications Includes:

Powered by ROAR Online Publication Software from Lions Light Corporation
© Copyright 2020

Rendered 10/30/2020 01:27