Created: April 11, 2020
Modified: April 11, 2020
Modified: April 11, 2020
teaching
This page is from my personal notes, and has not been specifically reviewed for public consumption. It might be incomplete, wrong, outdated, or stupid. Caveat lector.- Dave's principles of effective teaching.
- Motivation is by far the most important thing. A student who wants to learn will learn even with a bad teacher. Much of the job of a good teacher is providing motivation. Share a contagious love of the subject and help people understand why it's cool.
- One of the best sources of motivation is other students. If you're the only person learning a subject, you feel like maybe you're kind of weird or the subject's not that important. If you're surrounded by other people who want to learn the subject, all of a sudden there's now social validation for the importance of the learning. And there's potential for competitiveness.
- It's important to build solid conceptual scaffolding. In education, this means hearing the right things at the right time. I could (maybe) give you a great explanation of quantum mechanics in terms of linear algebra, but if you're not comfortable with linear algebra, it will have been wasted effort.
- Self-driven exploration is extremely powerful, partly because it aligns with motivation (you're more motivated to learn things you're curious about, than things someone else just tells you), and partly because it aligns with your conceptual scaffolding: the questions you choose to ask or explore will be those that you can represent with your current concepts, so the answers will be representable and not wasted.
- Still, you will sometimes need a push to add new scaffolding: the stuff you don't know that you don't know.
- It's much better to figure things out for yourself, than to be told. Even if you're told, you still need to figure things out for yourself in order to retain them. Lecture is no substitute for homework and projects. This basic idea is the core of the Socratic method: when a teacher asks questions that force the student to 'figure out' the subject for themselves---or at least, to highlight what the student doesn't know and therefore make them curious to rectify their ignorance---the student ends up feeling ownership of the answers. Having to speak and explain a subject forces you to build a generative model, which leads to confidence; reading and lecture alone tend to create discriminative models (generative vs discriminative modeling).
- To encourage self-driven exploration, the learning process needs a goal. 'I want to understand this paper' or 'I want to build a system that does X' are 'lighthouse on the horizon' points that motivate learning all the necessary things along the way. At the beginning, the student might not know enough to formulate precise goals. The teacher can help, but (especially for adults) ultimately the goals should come from the student. It's okay if the initial goal is vague; we can refine it later on.
- Spaced repetition is a powerful way to remember facts and to retain skills. If you solve for the eigenvalues of a matrix ten times in a day, that's great. If you do it again a couple times in a few weeks, and again in a few months, and again in a year, now we're talking. Every time you retrieve and reinforce the skill, it gets stronger. Once it's not in your short-term memory, you're forced to develop higher-level conceptual structure that will help you remember it. The first time you learned the skill, it might have seemed arbitrary and weird. After a month or a year of learning other things, when you revisit the original skill, you'll notice connections to other concepts you've learned; those connections will let you create a more parsimonious representation of the original skill.