By saying that procedural memories can be trained efficiently with mnemosyne, you are implying that the algorithm fits both types of memories, thereby implying that they share the same forgetting curve. That's your hypothesis, so the burden of proof is on you. I'm not being combative or anything, but that's how it is. Sufficiency. If I propose that bananas cause extreme diarrhea in Russian children under the age of 3 that live in remote villages untouched by tropical fruits, and there is no evidence to contradict me, would it seem reasonable for YOU to prove me wrong? Absurd? Yes, but it illustrates the point.
Either way, I think your idea is great and I hope that you or somebody more familiar with python can realize it. You mentioned math and Go as potentially suitable areas of application. What else do you think could realistically benefit from this? On Jan 7, 10:19 am, Gwern Branwen <[email protected]> wrote: > On Fri, Jan 7, 2011 at 12:04 AM, dazedconfused <[email protected]> > wrote: > > I'm wondering whether the same algorithm would be appropriate for > > procedural memory. I'm not intimately acquainted with Ebbinghaus' > > work, but I think it was based on declarative memory. These types of > > memory are executed by different parts of the brain (declarative in > > hippocampus, procedural is less understood), so there may indeed be > > differences. Any thoughts? Research? > > > I play piano, and if I don't practice a song from time to time, I > > forget how to play it completely. It's not that I can't remember the > > names of the chords, but my hands just don't know where to go. A > > mnemosyne card that reminds me to play a given song would be great, > > but I'm not sure that the algorithm would be appropriate for that sort > > of memory. > > If the user is honestly grading the cards (as we must always assume), > then either the cards will be scheduled too much or too little. If > they are scheduled too much, then the user will eventually get tired > and start marking them as 5, at which point they get blown away to > years intervals fairly soon. If they are scheduled too little, then > the user will fail each time they rarely come up, which will show up > in the statistics eventually and be diagnosable, proving the problem. > > And is there any real evidence to think that the forgetting curve for > more-procedural stuff has a different shape? If there isn't, I think > the burden is kind of on whomever wants to argue that it has some > different form. > > -- > gwernhttp://www.gwern.net -- You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/mnemosyne-proj-users?hl=en.
