Has this been considered?
Smart scheduling: AI algorithms can analyze the user’s learning patterns, memory retention, and historical performance data to optimize the scheduling of card reviews. This can help prioritize cards that need more frequent review and reduce review frequency for cards that are already well-learned.
Everyone would have their own custom scheduling, finally making it more efficient than SuperMemo.
I did use gpt4, but it cant do something.
if AI could do anything, you simply ask AI to guess WTF is SM-18,
then re-implement it as SMA-18, that will solve the problem.
oops, SMA is same as spinal muscular atrophy, may be use other names.
There is also this add-on; I have yet to try it, but it looks quite intriguing: https://ankiweb.net/shared/info/447942356.
From the github:
Tired of having to deal with anki flashcards that are too similar when grinding through your backlog? This python script creates filtered deck in optimal review order. It uses Machine Learning / AI to make semantically linked cards far from one another.
Here are different ways of looking at what AnnA can do for you in a few words:
- When you don’t have the time to complete all your daily reviews, use this to create a special filtered deck that makes sure you will only review the cards that are most different from the rest of your reviews.
- When you have too many learning cards and fear that some of them are too similar, use this to automatically review a subset of them.
- AnnA helps to avoid reviewing similar cards on the same day.
- AnnA allows to reduce the number of daily reviews while increasing (and not keeping the same) retention.
It’s AI, but not large model. It only has a dozen weights.
This seems to focus more on similar cards.
Algorithm SM-18 - supermemo.guru
One example is that SM18 takes into account that cards can randomly become more difficult over time. AI would be able to figure this out as well.