open-spaced-repetition / heterogeneous-memory-research

some manuscripts and notebooks about spaced repetition research
2 stars 0 forks source link

Card hashes? #1

Open Mathnerd314 opened 2 months ago

Mathnerd314 commented 2 months ago

A person can only memorize a card once at a given time, making multiple memorizations under identical conditions impossible. [...] Due to data heterogeneity, any memory model trained on review logs essentially estimates the mean memory states (stability, difficulty, and more).

While one person can only memorize/forget a card once, multiple people learning the same card might have the same patterns. This is already happening as there are many "shared decks" being re-used by lots of people. It seems worthwhile to determine if the stability / difficulty of a card are the same for everyone.

Although it seems card hashes are not in the Anki dataset, the Mnemosyne dataset includes card hashes for one - although it is not publicly available, perhaps you could obtain it where others have failed.

L-M-Sherlock commented 3 weeks ago

I guess the largest shared deck of Anki is the med deck of @AnKingMed's https://www.theanking.com/med-student

But I'm not sure whether they could share the data.

AnKingMed commented 3 weeks ago

We aren’t collecting data currently but are hoping to in the near future and then we can share it