At the AME Roundup Conference 2020 in Vancouver, BC, which finished on Jan. 23, 2020, a handful of enterprising technology-focused companies made their pitches on Tuesday, Jan. 21, 2020 in front of an audience of people who had heard much of the same thing the year before and were probably waiting for fresh revelations.
Stratus Aeronautics – Drones!
Two of the company representatives had straightforward hardware to offer: – The first, Stratus Aeronautics from Burnaby, B.C., represented by Curtis Mullan, offers long-distance, high-detail level, magnetic and aerial survey, inspections and photography using UAV (Unmanned Aeronautical Vehicles), i.e. drones. Drones is a hot topic these days. Lately their drone, the Venturer, which has military and industrial grade components, has demonstrated flights of up to 300 km in one day, at speeds of 60 to 70 km/hr. The resulting footage can be used for geogrammetry for 3D models of mine sites. They are overcoming problems caused by strong winds, sub-zero temperatures, the stability of the camera boom, noise appearance in the image and video signals, and electro-magnetic interference. (Ref.: Signal & Image Processing: An International Journal (SIPIJ) Vol.6, No.2, April 2015). Stratus claims that they have reduced interference to 0.04 nT (nanotesla).
Well, yes. Every so often when surveys have to be done I hear someone say: “Drone, perhaps?” But in every jurisdiction there is a plethora of laws and bylaws to deal with, including, in Canada, VTOL restrictions.
KORE Geosystems – Taking the schlepp out of core logging
Another company pitching its hardware was KORE Geosystems, represented by Chris Drielsma, its Senior Advisor. Newmont Goldcorp recently invested $1 million into the company – so they must have faith of some kind. KORE is based in Toronto and Melbourne, Australia. Basically KORE has digitized the labour-intensive job of core logging with their “SPECTOR” machine and system. (Gosh, doesn’t that name just remind you of a James Bond film?) SPECTOR is made up of SPECTOR instrumentation (SPECTOR Optics), SPECTOR software and SPECTOR AI. SPECTOR Optics is a machine into which you put a piece of drill core, and the machine then photographs it closely. According to the website;
“SPECTOR Optics is an efficient, intuitive core imaging system, which rapidly captures high-resolution photos. {…] SPECTOR Optics images are acquired with a resolution of 100 micro-meters per pixel.”
Ah, well, someone still has to do the drilling. And some experienced Geologist still has to look at the drill core and classify it.
SPECTOR Artificial Intelligence (AI) is still in the “learning” stage. It means that human input and analyses are still the basis of the models used.
“SPECTOR AI […] has models to segment rock, classify lithology and alteration, detect veins and localize fractures. All AI predictions are generated in the cloud after an image uploaded from SPECTOR Optics and can be visualized, accepted, or corrected in SPECTOR Geo. The AI models are regularly trained and with more data accumulating over time and corrections made by the geologist, SPECTOR AI learns and improves accuracy.”
Note that last phrase: “corrections made by the geologist”. To be sure, this would save time for laborious schlepp by the Geology team, but this product is still in development. As one client put it, “We look forward to KORE continuing to customize their product to support site-specific needs.” So, there is still some way to go before your average Geology team can rely only on the big black-and-yellow box.
The problem with data in a black box (of a different kind)
Whenever the use of data comes into the picture, whether rock core data or land surface data, it seems that tech companies come up against two prevalent problems:
The first problem is how they use the data: how they source, collect, homogenize, interpret and define the data, and how they convert those words and numbers from human to machine language.
The second problem is how to ensure the compatibility of the data to global standards and forms, for interoperability.
Saying you can “do AI”, meaning that you own computerized, probably internet-based, databases on which you can run queries that would otherwise take too long for a human to do, is conceivable. But in most cases, everything around the data that feeds into the machine learning program is a “black box”. Who knows what goes in, and who knows the value of what comes out? How is the data being manipulated? And where is it?
Note: A black box is a device, system or object which can be viewed in terms of its inputs and outputs (or transfer characteristics), without any knowledge of its internal workings.
Drielsma made the disingenuous comment that KORE “keeps digital archives of images in the cloud that will last forever.” Well, that isn’t true. For one, “the cloud” is not some permanent place. It is just a way of saying that another company will store your data somewhere outside of your computer, and that you can access it over the internet. Physical storage spans multiple servers (sometimes in multiple locations), and the physical environment is typically owned and managed by a hosting company. Cloud storage only lasts as long as the service provider does. Secondly, in data storage, there is no “forever”. There is such a thing as data degradation and yes, it does happen.
Minerva Intelligence – Opening up the black box of AI
The scientists of Minerva Intelligence, from Vancouver, B.C., have been working away at the problem of machine learning in the earth sciences for decades. In May 2019, the company listed on the TSX Venture Exchange and it has opened an office in Darmstadt, Germany. Clinton Smyth, the CTO of Minerva, pointed out when he took to the stage that Minerva is not a “tech company”, but a “knowledge engineering” company that provides solutions in the mining and geohazard industries.
Watching Smyth present a clean, simplified version of the advanced programming that his company does (he is the founder along with Prof. David Poole of UBC), I saw mostly incomprehension on the faces of those in the audience. Minerva is doing what other technology companies with claims of working in AI do not seem to do: they are working directly on the systems and processes to transform data, and they are working directly with institutions and regulators that are setting the standards for data interoperability. They are opening up the “black box” – you can try it one aspect of it yourself, on their platform.
Potato, potahto, tomato, tomahto – and apples
The problem is demonstrated by talking about apples. It’s like saying: OK, I define an apple as being this. But you define an apple as being something else, and you’re spelling it differently. And, oh dear, you are writing it 苹果 , or you, over there, are writing it яблоко – now what? Who’s right, and how do we ensure that if you do a search online for “apple”, you’ll get the right results, and all the results? Is there a standard way to define the concept of an apple and categorize it properly in Pomology?
Who cares about sharing knowledge?
If the definition of an apple has so many angles, imagine the complexity of the body of knowledge in Earth Sciences – which is where Minerva works. Most people don’t care whether what they know and the scientific language they use can be used by others. Every company works in a silo, each with their own databases which they guard as their confidential IP. Seequent, the makers of Leapfrog GEO 3D modelling software, released a report in 2019, “The Data Management Challenge”, which pretty much spells out what I’m saying here and what Smyth explained. (Download it here: Seequent-Data-Management-Report-March-2019) On their website, Minerva states:
“Minerva believes that the use of AI can revolutionize how companies search for mineral deposits. That’s exactly why we created TERRA, our AI platform for mineral exploration. With TERRA, users can standardize their mining and exploration data to ensure interoperability, enabling them to find, share and use that data for more sophisticated analytics by either a human or through machine learning and reasoning.”
Minerva is working with global organizations outside of Canada to ensure that the geoscience data that they base their solutions on meets global standards for format and interoperability:
“As data standardization is still in its infancy, Minerva is a strong supporter of and contributor to the development and maintenance of internationally curated vocabulary standards, such as the INSPIRE initiative in the European Union. Data interoperability standards for earth sciences established by INSPIRE address the problem of non-standard taxonomies. Minerva is actively working with INSPIRE to assist in creating data standards and is contributing to its improvement by identifying problems in the standards when the data is of insufficient quality for use with Minerva’s technology.”
In fact, Minerva won the INSPIRE Helsinki 2019 Data Challenge for its innovative practical uses of spatial data in the “Landslide Application for Veneto, Italy” project. And what do they do with that cleaned-up data?
Amongst other things, their mining clients use Minerva’s TERRA AI platform to “find the best locations for exploration, to explain in detail why each location was identified, and to provide advice on what additional exploration information to look for.” Also, they say they can let their clients “overcome resource limitations with hundreds of deposit-specific targets or increase [their] R.O.I. in drill hole assay data.” Now, those are compelling statements for any exploration company. Isn’t discovering the next big resource the holy grail of mining?
When you ask, Which standards to you adhere to?, an engineering firm’s people will probably be able to give you lists of design standards and codes, which they – proudly – stick to. But when you ask, Can your data be read by anyone on any system?, people think it’s an odd question. Minerva is facing these challenges like Jack the Giant Killer, but at this stage of the game, I’m not sure who’s going to win.
Further reading about the problem with data: Commission for the Management and Application of Geoscience Information.
Like this:
Like Loading...
Geoscience & Communications 
At the AME Roundup Conference 2020 in Vancouver, BC, which finished on Jan. 23, 2020, a handful of enterprising technology-focused companies made their pitches on Tuesday, Jan. 21, 2020 in front of an audience of people who had heard much of the same thing the year before and were probably waiting for fresh revelations.
Stratus Aeronautics – Drones!
Two of the company representatives had straightforward hardware to offer: – The first, Stratus Aeronautics from Burnaby, B.C., represented by Curtis Mullan, offers long-distance, high-detail level, magnetic and aerial survey, inspections and photography using UAV (Unmanned Aeronautical Vehicles), i.e. drones. Drones is a hot topic these days. Lately their drone, the Venturer, which has military and industrial grade components, has demonstrated flights of up to 300 km in one day, at speeds of 60 to 70 km/hr. The resulting footage can be used for geogrammetry for 3D models of mine sites. They are overcoming problems caused by strong winds, sub-zero temperatures, the stability of the camera boom, noise appearance in the image and video signals, and electro-magnetic interference. (Ref.: Signal & Image Processing: An International Journal (SIPIJ) Vol.6, No.2, April 2015). Stratus claims that they have reduced interference to 0.04 nT (nanotesla).
Well, yes. Every so often when surveys have to be done I hear someone say: “Drone, perhaps?” But in every jurisdiction there is a plethora of laws and bylaws to deal with, including, in Canada, VTOL restrictions.
KORE Geosystems – Taking the schlepp out of core logging
Another company pitching its hardware was KORE Geosystems, represented by Chris Drielsma, its Senior Advisor. Newmont Goldcorp recently invested $1 million into the company – so they must have faith of some kind. KORE is based in Toronto and Melbourne, Australia. Basically KORE has digitized the labour-intensive job of core logging with their “SPECTOR” machine and system. (Gosh, doesn’t that name just remind you of a James Bond film?) SPECTOR is made up of SPECTOR instrumentation (SPECTOR Optics), SPECTOR software and SPECTOR AI. SPECTOR Optics is a machine into which you put a piece of drill core, and the machine then photographs it closely. According to the website;
Ah, well, someone still has to do the drilling. And some experienced Geologist still has to look at the drill core and classify it.
SPECTOR Artificial Intelligence (AI) is still in the “learning” stage. It means that human input and analyses are still the basis of the models used.
Note that last phrase: “corrections made by the geologist”. To be sure, this would save time for laborious schlepp by the Geology team, but this product is still in development. As one client put it, “We look forward to KORE continuing to customize their product to support site-specific needs.” So, there is still some way to go before your average Geology team can rely only on the big black-and-yellow box.
The problem with data in a black box (of a different kind)
Whenever the use of data comes into the picture, whether rock core data or land surface data, it seems that tech companies come up against two prevalent problems:
The first problem is how they use the data: how they source, collect, homogenize, interpret and define the data, and how they convert those words and numbers from human to machine language.
The second problem is how to ensure the compatibility of the data to global standards and forms, for interoperability.
Saying you can “do AI”, meaning that you own computerized, probably internet-based, databases on which you can run queries that would otherwise take too long for a human to do, is conceivable. But in most cases, everything around the data that feeds into the machine learning program is a “black box”. Who knows what goes in, and who knows the value of what comes out? How is the data being manipulated? And where is it?
Note: A black box is a device, system or object which can be viewed in terms of its inputs and outputs (or transfer characteristics), without any knowledge of its internal workings.
Drielsma made the disingenuous comment that KORE “keeps digital archives of images in the cloud that will last forever.” Well, that isn’t true. For one, “the cloud” is not some permanent place. It is just a way of saying that another company will store your data somewhere outside of your computer, and that you can access it over the internet. Physical storage spans multiple servers (sometimes in multiple locations), and the physical environment is typically owned and managed by a hosting company. Cloud storage only lasts as long as the service provider does. Secondly, in data storage, there is no “forever”. There is such a thing as data degradation and yes, it does happen.
Minerva Intelligence – Opening up the black box of AI
The scientists of Minerva Intelligence, from Vancouver, B.C., have been working away at the problem of machine learning in the earth sciences for decades. In May 2019, the company listed on the TSX Venture Exchange and it has opened an office in Darmstadt, Germany. Clinton Smyth, the CTO of Minerva, pointed out when he took to the stage that Minerva is not a “tech company”, but a “knowledge engineering” company that provides solutions in the mining and geohazard industries.
Watching Smyth present a clean, simplified version of the advanced programming that his company does (he is the founder along with Prof. David Poole of UBC), I saw mostly incomprehension on the faces of those in the audience. Minerva is doing what other technology companies with claims of working in AI do not seem to do: they are working directly on the systems and processes to transform data, and they are working directly with institutions and regulators that are setting the standards for data interoperability. They are opening up the “black box” – you can try it one aspect of it yourself, on their platform.
Potato, potahto, tomato, tomahto – and apples
The problem is demonstrated by talking about apples. It’s like saying: OK, I define an apple as being this. But you define an apple as being something else, and you’re spelling it differently. And, oh dear, you are writing it 苹果 , or you, over there, are writing it яблоко – now what? Who’s right, and how do we ensure that if you do a search online for “apple”, you’ll get the right results, and all the results? Is there a standard way to define the concept of an apple and categorize it properly in Pomology?
Who cares about sharing knowledge?
If the definition of an apple has so many angles, imagine the complexity of the body of knowledge in Earth Sciences – which is where Minerva works. Most people don’t care whether what they know and the scientific language they use can be used by others. Every company works in a silo, each with their own databases which they guard as their confidential IP. Seequent, the makers of Leapfrog GEO 3D modelling software, released a report in 2019, “The Data Management Challenge”, which pretty much spells out what I’m saying here and what Smyth explained. (Download it here: Seequent-Data-Management-Report-March-2019) On their website, Minerva states:
Minerva is working with global organizations outside of Canada to ensure that the geoscience data that they base their solutions on meets global standards for format and interoperability:
In fact, Minerva won the INSPIRE Helsinki 2019 Data Challenge for its innovative practical uses of spatial data in the “Landslide Application for Veneto, Italy” project. And what do they do with that cleaned-up data?
Amongst other things, their mining clients use Minerva’s TERRA AI platform to “find the best locations for exploration, to explain in detail why each location was identified, and to provide advice on what additional exploration information to look for.” Also, they say they can let their clients “overcome resource limitations with hundreds of deposit-specific targets or increase [their] R.O.I. in drill hole assay data.” Now, those are compelling statements for any exploration company. Isn’t discovering the next big resource the holy grail of mining?
When you ask, Which standards to you adhere to?, an engineering firm’s people will probably be able to give you lists of design standards and codes, which they – proudly – stick to. But when you ask, Can your data be read by anyone on any system?, people think it’s an odd question. Minerva is facing these challenges like Jack the Giant Killer, but at this stage of the game, I’m not sure who’s going to win.
Further reading about the problem with data: Commission for the Management and Application of Geoscience Information.
Share this:
Like this: