Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Spaced repetition for efficient learning (2019)

gwern.net

Post reply on HN
Print view View original

31–40 of 62 posts

Re: Spaced repetition for efficient learning (2019)

#31

Post by vunderba »

criley2 wrote: post #15
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

I have a half dozen language learning apps on my phone and have vibe coded a few concepts as well and while spaced recognition is amazing, it still suffers from the duolingo "vocabulary is not a language" problem. IMO the way around users feeling like spaced recognition isn't progression is by redefining progression away from memorizing vocabulary into into becoming proficient in conversation both listening and speak…

I’ve pointed this out before on HN, but if you want to use ChatGPT as a language partner, you must provide a topic. Expecting it to behave as a proactive teacher is a recipe for disappointment.

Here’s what I typically do:

- Create a custom GPT (mine is called Polly the Glot) with a system prompt instructing it to act as a language partner that responds only in Chinese or your target language of choice. Further specify that the user will paste a story or topic before beginning practice, and that this should guide the discussion.

- Start a new chat.

- Paste in an article from AP/Reuters.

- Turn on Voice Chat.

At that point, I’ll head out to walk my dog and can usually get about 30 minutes to an hour of solid language practice in.

Fair warning, you'll likely need to be at least an intermediate student by this point otherwise it'll probably be too much over your head.

Caveat: You could including a markdown file of your known vocabulary as an knowledge attachment in the custom GPT but I've no idea how well that would work in practice.

Re: Spaced repetition for efficient learning (2019)

#32

Post by sxg »

zakki wrote: post #6
charcircuit wrote: post #3

>I like Mnemosyne (homepage) myself - Free, packaged for Ubuntu Linux It seems to use a 2 decade old modification of a now 4 decades old algorithm which will be worse and waste more of the user's time than using Anki with FSRS or SuperMemo with SM-18.

Do you have recommended apps using those algorithms?

Mochi.cards

Very similar to Anki but with a sane UI.

Re: Spaced repetition for efficient learning (2019)

#33

Post by OriginalPenguin »

boxed wrote: post #30
yard2010 wrote: post #4

Anecdotally I stumbled upon this phenomenon when trying to learn how to play the piano. I noticed that at the end of a session I make so many mistakes and feel like I didn't learn that much, but coming back to it after a day or two I really felt the difference.

Motor learning is quite different from the type of information the article talks about. I tried adding dance moves into Anki to do spaced repetition and it's extremely obvious that it's a great way to remember a move very badly but never getting good at it. Compare that to the geography deck where Anki is just perfectly suited for the task and smashes it.

Spaced repetition works well for motor learning. You just have to keep hitting “Again” until you are actually good at it.

Re: Spaced repetition for efficient learning (2019)

#34

Post by ChadNauseam »

SteveMqz wrote: post #7
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

Well the simplest thing would be to have test/benchmarks you do at the start and end of each “chapter” (however you define that). You fail miserably at the test at the start of each chapter, and crush it at the end. The difficult part is deciding how the tests are spread out.

I love that idea. Thanks. I'm going to try to implement that!

Re: Spaced repetition for efficient learning (2019)

#35

Post by ChadNauseam »

zahlman wrote: post #20
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

Just have the program point out the size of the word pool it's drawing from. "Number go up" is objective and clearly motivates people in other disciplines (someone else mentioned lifting weights).

I've just added this yesterday. It now shows you how close you are to knowing 95% of the words of various movies. I think movies are a good goal, but I want earlier milestones too, so I need to think more. (Maybe spongebob episodes would be good actually)

Re: Spaced repetition for efficient learning (2019)

#36

Post by ChadNauseam »

sodality2 wrote: post #9
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

The solution is to learn content that you actually use with some regularity in your life outside of the testing! If you're doing this for education, the payoff might be the exam; if you're doing it to learn things without some particular end goal, you'll have to make your own way to make it worth it. The language learning app people could try scheduling monthly video chats with native speakers (swapping turns halfway…

haha, that's a great point. I need to find more ways to do that. Maybe I should put $targetlanguage songs on my playlists so I'll get happy as I'm able to recognize more and more

Re: Spaced repetition for efficient learning (2019)

#37

Post by ChadNauseam »

ben_w wrote: post #8
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

> I'm not sure what the best way to solve this problem is. I would much appreciate any advice. Serious answer: All the dark patterns. Loot boxes even if they only give users a digital hat, small animated bird (like the green one, but not) doing a silly dance when users get enough correct answers, some weird phrases sprinkled amongst the lessons which make the users laugh. Just, please let them have an off switch for…

Something funny is that I watched someone use my app for the first time yesterday, and she specifically requested more dark patterns. She said "you should add something that makes you want to come back like the duolingo owl looking sad". I don't know if I can bring myself to do that, but I have a feeling that I can find a way to play to my strengths rather than copy duolingo. Specifically, my advantage is that my app will actually teach you the language, so I can maybe find a way to turn that into motivation rather than needing to rely on dark patterns.

Re: Spaced repetition for efficient learning (2019)

#38

Post by ChadNauseam »

michaelcampbell wrote: post #12
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

FSRS has dials to tweak to show you cards either more or less often, so if it's feeling "too hard" you can tweak that dial to show you cards when "easier" (at the risk of seeing more cards per day since you're seeing them more often). Failing to do that, one might consider that instead of focusing on how hard each card feels, but rather the size of the corpus that they have "under their belt". This is the case if you…

True. Changing the recall rate would definitely make it easier. But like you said, it increases the workload. I'm not sure that's the optimal play.

I want to find a way to show people how much they're learning. And maybe make it more fun. But what I really want to do is change their mindset from "wow, that was easy, i'm learning so much" to "wow, that was hard, i'm learning so much".

Re: Spaced repetition for efficient learning (2019)

#39

Post by ChadNauseam »

Denzel wrote: post #17
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

‘Desirable difficulty’ is the research term. To solve your problem, first understand your users need a mindset change. We need to connect their action to a “satisfying feeling” as you said. You want your users to be like weight lifters. No lifter comes out the gym saying, “Man that was the best workout, felt so easy,” to the contrary, lifters use progressive overload to induce difficulty because that difficulty conne…

That's a great analogy. I'll need to think about how to message that difficulty is good. It's a tricky proposition.

Re: Spaced repetition for efficient learning (2019)

#40

Post by ChadNauseam »

criley2 wrote: post #15
ChadNauseam wrote: post #5

One counterintuitive issue with spaced repetition is that the modern algorithms like FSRS are actually almost too good at scheduling. The effect is that each card is almost perfectly scheduled to be very difficult but still doable. Now, it's a bit weird to call it an issue considering it's the whole concept working exactly as designed. But it does cause one follow-up problem. The problem is that in life, we are accus…

I have a half dozen language learning apps on my phone and have vibe coded a few concepts as well and while spaced recognition is amazing, it still suffers from the duolingo "vocabulary is not a language" problem. IMO the way around users feeling like spaced recognition isn't progression is by redefining progression away from memorizing vocabulary into into becoming proficient in conversation both listening and speak…

If you've vibe coded a few language learning apps, maybe we should collab. Check my bio to see what I've done in this space. I'm trying to make the best free language-learning app on the planet
Post reply on HN