mnemosyne VS srs-benchmark

I wonder if there is plan for this to land in Mnemosyne[1]. I prefer Mnemosyne over Anki because I can self-host the web-sync server.

1: https://mnemosyne-proj.org/

I have tried using spaced repetition with Mnemosyne for math, specifically for learning Category Theory. It did help. Spaced repetition seems to work better for me if the answers to the questions are short (like learning Spanish vocabulary). When doing math, you often want to remember an entire definition, which might be too long to use spaced repetition flash cards effectively.

I've had good luck with spaced repetition using mnemosyne for lots of other stuff but haven't tried it for knots yet

Also, take the time to learn everyone's name and face. I use a flash card program like Mnemosyne to copy people's photos from the corporate directory. Learn them all the first week or even in the first couple of days.

https://mnemosyne-proj.org/ use this, make every relevant term a flash card and event a flash card. It takes forever to populate, but you learn on entry in addition to “study”.

[15] https://mnemosyne-proj.org/ - OS alternative to Anki

Anki is a spaced-repetition system based on Mnemosyne for studying flashcards in such a way that optimally loads facts into and maintains them in long-term memory. During daily study sessions, Anki shows flash cards, and you self-rate your ability to recall each correct answer. You review easier cards on a maintenance schedule, and Anki schedules the ones you have trouble with more frequently.

Some time ago I saw several posts about a machine learning based scheduler for Anki. There is another project known as mnemosyne that conducts research into long term learning. Do you want similar changes to Anki's scheduler?

I've shared it here in the original ticket that added the benchmark confidence intervals https://github.com/open-spaced-repetition/fsrs-benchmark/iss...

Here's the link to the benchmark repository: https://github.com/open-spaced-repetition/fsrs-benchmark/tree/Feat/new-dataset

Recently, I have run three comparisons for spaced repetition algorithms. They included SM-15, SM-17, SM-2, HLR, LSTM and FSRS series. The initial result shows that FSRS v4 beats all other algorithms in predicting probability of recall. It's a good news that the open-source algorithm can overperform SuperMemo's proprietary algorithm.