Investing in Data Talk
Download as PPTX, PDF0 likes58 views
The document discusses the challenges and considerations of implementing machine learning in mining, emphasizing that it is not a silver bullet but a tool that requires expertise in various fields, including data science and management. It highlights the importance of focusing on results, behavior change, and obtaining useful data rather than just running pilots. The author argues that improving productivity, profit, and safety should be the ultimate goals, with data being the only significant asset in this context.
Investing in Data Talk
- 3. Anyone consistently making money from Machine Learning in Mining, not including consulting?
- 8. Machine learning is a tool… You have to figure out how to bring it into your workplace…
- 10. Computation Data Program Results Computation Data Results Program Machine Learning Makes Software Development Different
- 11. Machine learning starts with results It automates complex situations where rules are hard to understand
- 12. Change Behaviour Decide Analyse Value Get Useful Data TIME - How fast does the cycle need to happen to get value? SCALE - Are we going to diversify or replicate? AUTOMATION - How automated is the process? FEEDBACK - How is improvement going to happen? Value is Complex…
- 13. Value Change Behaviour Decide Analyse Get Useful Data How hard is it? Very difficult Straightforward Where most of investment goes? Low High And most companies are doing it all wrong…
- 14. Machine learning pilots are easy compared to getting useful data and changing behaviour, or putting them in production and getting consistent value
- 15. Maturity models are an easy way to sell you consulting…
- 17. Machine Learning maturity is never a goal… Productivity, profit, safety — those are goals…
- 18. High Model Complexity Low Interpretability Required Amount of Training Data Model Performance Machine Learning models are not a silver bullet for your lack of vision or knowledge…
- 19. Change Behaviour Decide Analyse Value Get Useful Data What do you have to be good at… I.T. Applied Data Science Safety Geology Mine Engineering Maintenance Transportation Supply Chains … and so on Management
- 20. Machine Learning requires you to be good at mining and I.T. and Data Science and Management
- 21. Machine Learning in Mining… Machine Learning is a tool… starts with results… pilots are easy… maturity is never a goal… requires you to be good at mining…
- 22. The only asset in this story is Data…