Hacker News Reader

Comment by boredemployee | Hacker News Reader

any other recommendation for articles regarding techniques that works for most people?

This is pretty much the current state of the art in learning research: https://bjorklab.psych.ucla.edu/research/

From a more applied angle, a book like "10 steps to complex learning" might be helpful.

I come from a similar cog psych background as the Bjork Lab, so am a big fan of their research, but books like 10 steps come from instructional design, which is a bit more focused on the big picture (designing a whole course vs individual mechanisms).

Thanks for the tip.

This is good stuff but I'll say that it isn't as comprehensive as all that. These studies and findings are almost entirely focused on simple recall knowledge (isolated facts, vocabulary). That's an important part of learning certainly and it informs research on learning higher order concepts but it's not the full story.

Just to name one example, folks might look into research on conceptual change theory (eg Chi or Posner). This theory helps explain why a concept like electricity is so challenging to learn. The reason, in brief, being that naive conceptions make a category error and think of electricity as a thing rather than a process. And this theory then informs instructional practice. Specifically, teachers should be aware of difficult concepts and should design activities that force students to confront the contradictions between their naive models and more accurate/complete ones.

Mickie Chi also has fascinations research on active learning (ICAP) and related work on the effectiveness of peer learning.

Thanks for your example of the category error here. Are there other cases though? For example, sometimes I doubt my understanding of gradient descent because it is very hard to implement in code on data and show error reducing over time (writing from scratch). But in some examples of a few nodes I can calculate it perfectly. I can do the Maths manually.

What error might I be making? For the future, is there a list of types of errors?

Scott Young has a lot on this, including going deep on the literature and explaining it to populist student audience:

https://www.scotthyoung.com/blog/the-best-articles-on-learni...

For instance, on memory:

https://www.scotthyoung.com/blog/2019/02/15/memory/

The default algorithm in Anki, in case you don't know it

Or not, since they recently added FSRS as an alternative.

You're both right - SM2 is still the default in Anki.

It is super unlikely that an algorithm that was made based on feels would turn out to be the most effective. Especially since it requires some learning and errors to even learn how to use it.

Dunlosky J, Rawson KA, Marsh EJ, Nathan MJ, Willingham DT. Improving Students' Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology. Psychol Sci Public Interest. 2013 Jan;14(1):4-58. doi: 10.1177/1529100612453266. PMID: 26173288.

https://www.goodreads.com/book/show/19288640-how-we-learn

The comment I wrote below is weird. What I mean to say is google for articles on writing good prompts. The rest matters less.

-----

Ignore this:

Spaced repetition, if you can stick with it, works. The details around schedules etc. don't matter as much as doing it in a way that you can keep up.

Also writing good prompts is important. I have yet to write more extensively about this but here is a start: https://news.ycombinator.com/item?id=41126734

FSRS v5 is state of the art

Reply on news.ycombinator.com