sourceType:: book author:: Scott H Young sourcePublication:: Ultralearning ref:: noteTitle:: Scott H Young, Ultralearning (Book)
Scott H Young, Ultralearning (Book)
Principle 1: Metalearning.
Do metalearning research before and during the ultralearning project
Principle 2: Focus.
Keep Your Knife Sharp - Build focus, concentration, dedication.
Principle 3: Directness.
Learning activities done in connection with the context in which the skills will be used. This is often harder at first. Like having conversations in a new language vs. choosing translations from multiple choice.
Transfer: Learning something in one context and applying it in another.
Over and over, it has been shown that transfer rarely happens well. This is seen in schooling, but also in corporate training.
Transfer is narrow. ie. Latin and Geometry don’t help you “think better” in general (Thorndike & Woodward, 1901). “formal discipline theory” has basically been debunked.
Transfer is generally harder when knowledge is limited. More knowledge and skill in an area == more flexible and applicable outside contexts in which learned.
Most formal learning is very indirect.
Overcoming Transfer with Directness.
- Need for transfer is reduced with Directness.
- Directness may help with transfer to new situations because real-life situations share details (more so than with classrooms).
Learning Directly
- Project-based. (For purely intellectual pursuits, consider writing a thesis - an original paper)
- Immersive. ie. language immersion or for non-language, consider joining a community - like an open-source project for a programming learner.
- Flight simulator method. Simulate the environ to a degree to which the cognitive elements remain faithful.
- Overkill approach. Put yourself into an extremely demanding environment. You’ll be unlikely to miss feedback, given the high demands and risk. Aim for a challenge above the level that is strictly needed.
What skills do I want and where do I want to apply them?
If I attempt an Ultraleaning project, I’ll need that ansser. It’s useful outside that context as well, obviously.
I was pondering that initial question and realized “I want to learn more machine learning!” does NOT fit the bill. Why? to do what with? umm “To do beneficial research!” – but… that might not even be the right fore-knowledge. And to do Direct leanring I’d need to know what the “workday” there looks like. How do I find that out?
Principle 4: Drill.
Socratic method: probing questions rather than direct contradiction.
rate-determining step in chemistry: the slowest step in the process - ultimately determines the entire timing. bottleneck. identify the rate-determining step in your learning and drill it.
The true benefit of Ultralearning is Sequential ultralearning projects create a virtuous cycle