Sign up to save your podcasts
Or
Share The Artificial Geologist: Using Machine Learning & Neural Networks to Find Gold
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
The Artificial Geologist: Using Machine Learning & Neural Networks to Find Gold
The "motherlode" is just a model away. In this episode, Dr. Jeffrey Zurek (P.Geo) bridges between the classroom & the boardroom with colleague & data scientist Frederick Jackson from Computational Geosciences Inc. Together, they explore how machine learning (ML) & artificial intelligence are revolutionizing mineral exploration.
We dig into the expensive reality of drilling—where a single hole can cost over $100k—& how neural networks act as an "artificial geologist" to find patterns in massive "data cubes." From the Yilgarn Craton of Western Australia to the surprising links between finding gold, & detecting brain tumors, this episode proves that while the intelligence might be artificial, the discovery is real.
Topics Covered
- The Business of Discovery: How science functions in the corporate world & the real-world consequences of being wrong.
- Drilling by the Numbers: Why de-risking drill holes is the primary driver for AI in mining.
- The Data Cube: Integrating geophysics, radiometrics, etc., to build "geological ChatGPT."
- Neural Networks 101: Moving beyond simple regressions to non-linear, brain-inspired algorithms.
- Prospectivity & Policy: How heat maps help inform land-use decisions .
- Bioacoustics: Whimsical detour to tracking whales for conservation using the same ML technology.
Episode Chapters
(00:00) Intro: Geology Meets Algorithms
(02:05) The High Cost of Drilling: Why We Need Models
(04:35) Frederick Jackson Spinosaurus to Data Science
(07:50) Industry vs Academia: The Cost of Being Wrong
(10:10) The SEG Paper: Gold Prospectivity in Australia
(11:50) AI Hallucinations in Geophysics Managing Risk
(15:15) Building the Data Cube: Features vs. Labels
(19:35) Garbage In, Garbage Out: AI Pitfalls
(21:20) Neural Networks: an "Artificial Geologist"
(25:10) Results: Heat Maps and 2D De-risking
(30:45) Beyond Minerals: Tracking Mosquitoes & Brain Tumors
(32:45) Bioacoustics: Citizen Science & Whales
(34:30) The infamous Science Joke
Links & Resources
- Citizen Science: Orca Sound
- Support: Pateron
- Socials: Bluesky | Instagram | Facebook
Whimsical Wavelengths: Deep-dive conversations where a working scientist unpacks how we know what we know, one paper, one idea, or whimsical detour at a time. Hosted by Dr. Jeffrey Zurek (P.Geo).
View all episodes
By Jeffrey Mark Zurek - PhD, PGeo, Geophysicist Volcanologist Science CommunicatorThe "motherlode" is just a model away. In this episode, Dr. Jeffrey Zurek (P.Geo) bridges between the classroom & the boardroom with colleague & data scientist Frederick Jackson from Computational Geosciences Inc. Together, they explore how machine learning (ML) & artificial intelligence are revolutionizing mineral exploration.
We dig into the expensive reality of drilling—where a single hole can cost over $100k—& how neural networks act as an "artificial geologist" to find patterns in massive "data cubes." From the Yilgarn Craton of Western Australia to the surprising links between finding gold, & detecting brain tumors, this episode proves that while the intelligence might be artificial, the discovery is real.
Topics Covered
- The Business of Discovery: How science functions in the corporate world & the real-world consequences of being wrong.
- Drilling by the Numbers: Why de-risking drill holes is the primary driver for AI in mining.
- The Data Cube: Integrating geophysics, radiometrics, etc., to build "geological ChatGPT."
- Neural Networks 101: Moving beyond simple regressions to non-linear, brain-inspired algorithms.
- Prospectivity & Policy: How heat maps help inform land-use decisions .
- Bioacoustics: Whimsical detour to tracking whales for conservation using the same ML technology.
Episode Chapters
(00:00) Intro: Geology Meets Algorithms
(02:05) The High Cost of Drilling: Why We Need Models
(04:35) Frederick Jackson Spinosaurus to Data Science
(07:50) Industry vs Academia: The Cost of Being Wrong
(10:10) The SEG Paper: Gold Prospectivity in Australia
(11:50) AI Hallucinations in Geophysics Managing Risk
(15:15) Building the Data Cube: Features vs. Labels
(19:35) Garbage In, Garbage Out: AI Pitfalls
(21:20) Neural Networks: an "Artificial Geologist"
(25:10) Results: Heat Maps and 2D De-risking
(30:45) Beyond Minerals: Tracking Mosquitoes & Brain Tumors
(32:45) Bioacoustics: Citizen Science & Whales
(34:30) The infamous Science Joke
Links & Resources
- Citizen Science: Orca Sound
- Support: Pateron
- Socials: Bluesky | Instagram | Facebook
Whimsical Wavelengths: Deep-dive conversations where a working scientist unpacks how we know what we know, one paper, one idea, or whimsical detour at a time. Hosted by Dr. Jeffrey Zurek (P.Geo).