When reviewing Anki cards I have a tendency to memorize the visual look of a question. I then use that as a hook to recall the answer from memory. The problem is that in real life, I will almost surely never be presented the question asked in the same way as in Anki.
-
-
Prikaži ovu nit
-
The hook needs to be in the meaning of the question. In my opinion, what I just described is one of the biggest open problems in spaced repetition systems today.
Prikaži ovu nit -
A first step in solving this problem is paraphrasing the question. On each review for the same card you get asked questions that are semantically equivalent questions but written differently. Paraphrasing helps but you are still seeing the concept from only one perspective.
Prikaži ovu nit -
A second step is asking slightly (semantically) different questions about the same concept (how does it relate with [other concept]? + backwards, details, what is it useful for? etc.) In picturespic.twitter.com/mzyj67qDNX
Prikaži ovu nit -
I'm only talking about flashcards' questions because that's what matters. Questions are pointers to memory, question quality is essential for spaced repetitions systems to be effective. Answers just serve as reference.
Prikaži ovu nit -
I think
@michael_nielsen and@andy_matuschak will soon talk about this problem.https://twitter.com/michael_nielsen/status/1217900290369699840 …Prikaži ovu nit -
@gwern also touched on this. On his popular spaced repetition article https://www.gwern.net/Spaced-repetition … he talks about "dynamic SRS decks".pic.twitter.com/S5GAbb9v25
Prikaži ovu nit -
Me and
@vccr are building a spaced repetition system; we will test other ideas that go even further in solving the problem of real-life recalling of concepts.Prikaži ovu nit
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.
Thinking about SRS aka memory protheses.