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

Anki: Memorization with Spaced Learning

apps.ankiweb.net

Post reply on HN
Print view View original

61–70 of 175 posts

Re: Anki: Memorization with Spaced Learning

#61

Post by ainar-g »

jacekm wrote: post #35
ainar-g wrote: post #21

For anyone who wants to use spaced repetition for learning languages, I recommend Clozemaster[1]. It uses the cloze deletion technique and has thousands of sentences for most languages. Using it has significantly improved my German. [1] https://www.clozemaster.com/ [2] https://en.wikipedia.org/wiki/Cloze_test

Interesting, I just tried it for Japanese. But too often I could fill in a word without understanding it (there's often only one grammatical form that fits). Also each sentence is accompanied by English translation which feels a little bit like cheating ;) Does the method still works even with the translation available?

It works, but you can actually disable the translation in the settings, or make it show after you've answered.

Re: Anki: Memorization with Spaced Learning

#62

Post by nathanasmith »

inetsee wrote: post #51

I really like the idea of spaced repetition learning, but it's the little details that are driving me crazy. I cannot get Anki working on Raspberry Pi. (I have a Raspberry Pi connected to a TV in front of my rowing machine so I can study while I'm exercising.) I tried Mnemosyne but I can't get the sync server working, and I'm disappointed in the formatting options provided. I tried org-drill but adding images to card…

Have you tried just going to the website https://ankiweb.net/decks/ and reviewing your cards there? It should work on any device with a browser including your RasberryPi.

Re: Anki: Memorization with Spaced Learning

#64

Post by aasasd »

Noumenon72 wrote: post #26

I used Anki for about a year, and it definitely didn't work for me. Going through and remembering the answers was always a struggle. I only ended up with 36 cards -- the idea of making a bunch for every person I meet and book I read would overload my memory like crazy. My memory has always been really good at indexing "what's the word for that", "there is an algorithm for this called X", "the keyboard shortcut for th…

> Going through and remembering the answers was always a struggle.

That's a definite hint to adjust the procedure, e.g. change the schedule. There are plenty of variables: some cards are better learned together because they are too similar, and you'd have trouble recalling them among other ones (I have this when going over geography, with countries that are in busy regions like Oceania or the Caribbean) ― you can pull them into a "filtered deck" by a tag and study them separately. Or, you do the opposite and put several decks into a grouping deck so you have to remember different kinds of facts in mixture instead of drills on specific topics.

Some answers are poor fit for one's memory, so you change them until you feel comfortable and retain the knowledge. E.g. I can't remember population of countries and cities with good precision, so I allow myself some slack with the answers.

Regarding the schedule, I found that I have to have a tighter schedule than the default, and I also employ the "bury" feature to repeat some cards. Plus, I have different limits on the numbers of new and reviewed cards for different decks.

If you find that you need more context to retain the knowledge, I'd think it suggests that you should have more cards, not fewer, and they should have closely related information, even partially duplicating each other. Maybe also have more info on each card, such as code examples. It can be time-consuming to create the cards yourself, but there are ready-made decks, including plenty on programming, and it's also possible to convert cards made for other apps. (Might be also fruitful to see how others make cards, and borrow some ideas.)

Generally, IMO spaced repetition simply "automates" the natural need of the memory to have the facts repeated to remember them―and other tricks, such as addition of more context, would also help. But, of course, for some people the right approach might be to abandon flash cards altogether and employ something else, such as practical exercises.

Re: Anki: Memorization with Spaced Learning

#65

Post by graeme »

irq11 wrote: post #55
graeme wrote: post #5

I’m using anki to gradually learn japanese kanji using the RTK method. Working really well so far, up to about 800 with little struggle.

Take my advice, as a somewhat advanced learner: don’t do this. RTK is this weird honeypot for beginning learners, but it’s nearly always a mistake. The “discouraged RTK user” is almost a comic meme in Japanese language study. You will spend months “learning kanji”, only to discover that knowing kanji meanings, in isolation, is unhelpful in most situations. You can get all of the same benefits, with more immediate uti…

Thanks. So, I'm actually doing it as a side thing while learning German. I have found I can't do two languages at once, but that learning the characters doesn't interfere.

I should finish them by April or so, and be advanced enough in German at that point that I can switch to Japanese. At which point I'll do Pimsleur and Assimil.

I think having the character base will accelerate that process. I previously tried Pimsleur Japanese, but found it didn't go smoothly compared to European languages I've done. The lack of characters was a big hurdle.

Thanks though. Is discouraged RTK learner someone who completed RTK, or someone who fails to

Re: Anki: Memorization with Spaced Learning

#66

Post by nathanasmith »

jhayward wrote: post #50
vedtopkar wrote: post #48

Anki has become the de-facto standard for memorizing the large swaths of information required during medical school. Not everyone uses it, but pretty much everyone that does and sticks to it sees impressive results.

> Not everyone uses it, but pretty much everyone that does and sticks to it sees impressive results. This is another way of say "everyone that it works for gets good results"

It's interesting to note in studies comparing spaced repetition vs. massed learning, when students are asked afterward which method works better, they more often choose massed even though their own test results favor spaced repetition[0]. The upshot is people can use spaced repetition, it can "work" (in the sense that they remember the material better and test better) yet still drop it thinking massed practice to be superior.

[0]https://www.gwern.net/docs/spacedrepetition/2001-simon.pdf

Re: Anki: Memorization with Spaced Learning

#67

Post by DarthMader »

xkfm wrote: post #39
Marorin wrote: post #18

I always kinda wondered how to make Anki stuff for like algorithms and the like.

For Anki, I'd grab the complete article from Wikipedia/textbook/[good source], and then write a few examples. The flashcards would be a number of cloze deletions, and then I'd have a separate deck with a practice question to recall from a blank state. For example [1], I'd take this Wikipedia article, and break it down like so: In computer science, a stack is an abstract data type that serves as a collection of elemen…

That seems like a waste of flashcards to me, but if it works for you, it works. I would just create one flashcard for front and back. What is a stack in CS? A stack is an abstract data type that's a collection of elements, with two principal operations: pop and push. If an answer's longer than 1-2 sentences then it's time to break it apart into different questions.

Re: Anki: Memorization with Spaced Learning

#68

Post by irq11 »

graeme wrote: post #65
irq11 wrote: post #55

Earlier quoted context omitted.

Take my advice, as a somewhat advanced learner: don’t do this. RTK is this weird honeypot for beginning learners, but it’s nearly always a mistake. The “discouraged RTK user” is almost a comic meme in Japanese language study. You will spend months “learning kanji”, only to discover that knowing kanji meanings, in isolation, is unhelpful in most situations. You can get all of the same benefits, with more immediate uti…

Thanks. So, I'm actually doing it as a side thing while learning German. I have found I can't do two languages at once, but that learning the characters doesn't interfere. I should finish them by April or so, and be advanced enough in German at that point that I can switch to Japanese. At which point I'll do Pimsleur and Assimil. I think having the character base will accelerate that process. I previously tried Pimsl…

Someone who fails. The usual pattern is: new learner goes all-in on RtK, makes it a month or two, then gives up after realizing that they’re making no discernible progress, even though they “know” lots of kanji.

There are useful parts of RtK, but they’re the parts Heisig didn’t invent: radicals and mnemonics. You can (and should) employ both, but in the context of words. As you learn words, the “meanings” of kanji will naturally settle in your brain.

RtK people love to argue that it helps them infer the meanings of words and discriminate similar kanji. But in practice, it’s insanely difficult to do the former (except for simple words), and the latter is rarely necessary. It’s far more common to encounter 2-3 kanji compounds with unique meanings that recombine into words, than it is to find words that compose meanings from their individual kanji.

Re: Anki: Memorization with Spaced Learning

#69

Post by AlphaWeaver »

baxtr wrote: post #57
AlphaWeaver wrote: post #2

I started using Anki a few weeks ago when the SuperMemo article popped up on HN. I really like it, and it has helped me remember key terms and learn a new language, all in a really short time... Feel free to ask me questions if you're confused about how it works! I'm fairly experienced now.

I want to learn Spanish. I have a basic vocabulary and understanding. What’s the fastest way forward in your opinion?

On AnkiWeb there are a ton of decks available made by other people, including some really high-quality ones for Spanish. I'd recommend you download one of those, and then if there are words/content you're already familiar with, I believe you can go and set that you've already "learned" those words.

Re: Anki: Memorization with Spaced Learning

#70

Post by aasasd »

wodenokoto wrote: post #49
jacekm wrote: post #33

I've been using Anki for some time and I love it but it's not a perfect tool. What I especially like is it's flexibility when it comes to designing flashcards. You can adjust colors, font-sizes and pretty much every visual aspect using css code. You can even embed your own font within a deck, which can come in handy when you are dealing with Japanese kanji (Printed and handwritten kanji sometimes differ and iPhone co…

> Another problem (with the algorithm, not Anki itself) My main gripe with the algorithm is you can't take weekends off. I used Anki a lot in College, and Mondays where terrible, due to the backlog built-up over the weekend. One could also say, one shouldn't be drinking the weekend away ...

Actually, it's mentioned several times in the docs that you shouldn't think that backlog build-up means you have to crunch―or more specifically, IMO, 'backlog' isn't a thing here at all. If you can only put in thirty minutes a day, then do it and then close the app, this doesn't change because of a missed day. The cards will be there the next day, it's not like they catch fire when you skip a session. Stressing over the backlog would likely only make your mind work worse.

Instead, change the limit of cards to review to the number that you can get through in the time you have.

(Unless your college schedule requires you to actually crunch some number of cards every day? I don't know if that's a thing, but using spaced repetition in such a case would seem ill-fitting anyway.)

Post reply on HN