North of 60 Mining News - The mining newspaper for Alaska and Canada's North

Pebble Partnership copper gold molybdenum mine project Alaska Northern Dynasty NAK NDM

By Shane Lasley
Mining News 

Auryn to drill AI targets at Committee Bay

Drills to test first targets turned up with machine learning

 

Last updated 6/28/2019 at 4:34am

Artificial Intelligence AI machine learning generated gold drill targets

Daniel Murray; courtesy of Auryn Resources Inc.

Auryn Resources has applied machine learning to help narrow down drill targets in a highly prospective area surrounding the Three Bluffs gold deposit at the heart of its expansive Committee Bay property in Nunavut.

Auryn Resources Inc. is gearing up test intriguing drill targets that artificial intelligence has generated on the Committee Bay gold project in Nunavut.

"We are looking forward to turning the drills again at Committee Bay and to offer our shareholders the possibility of a major high-grade gold discovery using cutting-edge artificial intelligence applications," said Auryn Resources Executive Chairman Ivan Bebek.

At roughly 390,000 hectares (964,000 acres) Committee Bay is a massive property that covers a gold enriched greenstone belt that extends for roughly 300 kilometers (180 miles) across Nunavut's Kitikmeot region.

At the heart of this belt lies Three Bluffs, where 2.1 million metric tons of indicated resource averaging 7.85 grams per metric ton (524,000 ounces) gold; and 2.9 million metric tons of inferred resource averaging 7.64 g/t (720,000 oz) gold has been outlined.

Over the past couple of years, Auryn has identified two promising targets – Aiviq and Kalulik – along a 20-kilometer- (12.5 miles) long shear zone north of Three Bluffs.

Aiviq, located about 12 kilometers (7.5 miles) north of Three Bluffs, was first drilled by Auryn in 2017. One hole drilled that year, 17RGR003, cut 12.2 meters of 4.7 g/t gold, including 3.05 meters of 18.09 g/t gold.

The company drilled another 16 holes at Aiviq during 2018. The best hole of this program cut 13.5 meters of 1.54 g/t gold, including six meters of 3.3 g/t gold. This hole, 18RG012, was drilled toward the southwest end of the 2018 program.

Kalulik, situated about 10 kilometers (six miles) southwest of Aiviq, was also drilled last year. This drilling at Kalulik cut 10- to 20-meter-thick zones of low-grade gold mineralization. Results from this drilling include 21.34 meters at 0.4 g/t gold and 16.76 meters of 0.45 g/t gold.

While this drilling has yet to turn up a deposit with the size and grade Auryn is looking for, there are indications that such a deposit is hidden below the glacial till that masks the underlying geology at Committee Bay.

"We feel these systems (Aiviq and Kalulik) have the potential to be economic based on the geological characteristics observed in the drilling," said Auryn Resources Chief Geologist Michael Henrichsen.

To help refine drill targets across the "Three Bluffs playing field" – a 1,600-square-kilometer- (620 square miles) area which includes the Three Bluffs deposit and the shear zone that hosts Aiviq and Kalulik.

This avant-garde step of applying artificial intelligence to identifying drill targets was assisted by Computational Geosciences Inc., a company that utilizes advanced software tools for data processing, inversion, interpretation, and targeting.

These tools include VNet, a type of deep neural network that can handle an arbitrary number of 2D or 3D geoscience data inputs.

This style of deep learning for mineral exploration is an emerging technology that requires expertise in geoscience data processing, data interpretation, and artificial intelligence.

Auryn said the biggest advantage of a data-driven solution is to extract subtle correlations across multiple datasets over a large spatial area, all while reducing human bias.

For machine learning to work, however, accurate geological, geochemical and geophysical inputs are required.

"Simply put, garbage in equals garbage out," said Henrichsen. "However, in Auryn's case, our discipline to collect rigorous high-quality datasets allows for high-quality outputs from the machine learning."

VNet was "trained" to correlate the geological, geochemical and geophysical inputs to gold-rich drill intercepts previously encountered, primarily at Three Bluffs. This training was refined until machine learning could predict 99 percent of drill holes that contain gold mineralization.

"The patterns within the data that accurately predict 99 percent of drill holes is then applied across the area of study to derive the machine learning targets," Henrichsen explains in a video.

The AI program generated 12 targets in the study area, many of which correlated with targets previously identified by Auryn's technical team. This increases the team's confidence in both the machine learning and their own targeting.

Machine learning identified four target areas between Aiviq and Kalulik, including two targets immediately adjacent to Auryn's drilling; two Three Bluffs extension targets masked by glacial till cover; and identified targets along a 15-kilometer (nine miles) shear zone north of Aiviq.

Artificial Intelligence AI machine learning generated gold exploration map

Auryn Resources Inc.

Together, this provides Auryn with three parallel structural corridors to explore within the region where the machine learning was applied.

In addition, the AI program identified targets under lakes and areas where glacial sediments have completely masked any geochemical indications on the surface.

"This provides the technical team with targets that may otherwise have been overlooked and opens up a number of areas for further evaluation," said Auryn's chief geologist.

The roughly 3,000-meter drill program slated for Committee Bay this year will follow-up on four of the machine learning targets generated along the Aiviq-Kalulik corridor. To enhance the targeting effort, Auryn will carry out induced polarization geophysical surveys along this structural zone.

 

Reader Comments
(0)

 
 

Powered by ROAR Online Publication Software from Lions Light Corporation
© Copyright 2019

Rendered 07/18/2019 04:05