Joss Delage

Memory / SuperMemo / Piotr Wozniak

41 posts in this topic

All,

I'm reading an article on Wired that I find interesting and I'd love to get opinions from the members of this forum. The article is here:

http://www.wired.com/medtech/health/magazi...currentPage=all

Here are a few chosen quotes:

SuperMemo is based on the insight that there is an ideal moment to practice what you've learned. Practice too soon and you waste your time. Practice too late and you've forgotten the material and have to relearn it.

...

Fortunately, human forgetting follows a pattern. We forget exponentially. A graph of our likelihood of getting the correct answer on a quiz sweeps quickly downward over time and then levels off. This pattern has long been known to cognitive psychology, but it has been difficult to put to practical use. It's too complex for us to employ with our naked brains.

...

Twenty years ago, Wozniak realized that computers could easily calculate the moment of forgetting if he could discover the right algorithm. SuperMemo is the result of his research. It predicts the future state of a person's memory and schedules information reviews at the optimal time.

...

[Wozniak] wants to avoid random interruptions to a long-running experiment he's conducting on himself. Wozniak is a kind of algorithmic man. He's exploring what it's like to live in strict obedience to reason. On first encounter, he appears to be one of the happiest people I've ever met.

...

Wozniak takes an almost physical pleasure in reason. He loves to discuss things with people, to get insight into their personalities, and to give them advice...

...

From the business side of SuperMemo, Wozniak's priorities can sometimes look selfish. Janusz Murakowski, one of Wozniak's friends who worked as a manager at the company during its infancy, thinks that Wozniak's focus on his own learning has stunted the development of his invention. "Piotr writes this software for himself," says Murakowski, now a professor of electrical engineering at the University of Delaware.

...

When it comes to increasing intelligence, our brain is up to the task and our technology is up to the task. The problem lies in our temperament.

...

I find myself thinking of a checklist Wozniak wrote a few years ago describing how to become a genius. His advice was straightforward yet strangely terrible: You must clarify your goals, gain knowledge through spaced repetition, preserve health, work steadily, minimize stress, refuse interruption, and never resist sleep when tired. This should lead to radically improved intelligence and creativity. The only cost: turning your back on every convention of social life.

Share this post


Link to post
Share on other sites

Just wanted to add to my earlier post.

The brain is estimated to have one hundred billion neurons (nerve cells) that send the chemical messages that run our brains and influence our boidies. It is also estimated that there are more neurons in the average human brain than there are stars in the universe. And that some of these neurons make more than ten thousand individual connections with other neurons. So, please explain to me how one can easily estimate when exactly an individual will forget what they have been taught and then add in the fact that individual motivation on the subject under study has a lot to do with learning also.

Share this post


Link to post
Share on other sites
Just wanted to add to my earlier post.

The brain is estimated to have one hundred billion neurons (nerve cells) that send the chemical messages that run our brains and influence our boidies. It is also estimated that there are more neurons in the average human brain than there are stars in the universe. And that some of these neurons make more than ten thousand individual connections with other neurons. So, please explain to me how one can easily estimate when exactly an individual will forget what they have been taught and then add in the fact that individual motivation on the subject under study has a lot to do with learning also.

It is based on the forgetting curve by Ebbinghaus and experimental results. Spaced repetition really has nothing to do with reverse engineering the brain from the neurons on up.

I've used spaced repetition to become much more efficient at studying Korean. There are thousands of people using spaced repetition computer flashcard programs for learning languages now who are out pacing others by miles.

Share this post


Link to post
Share on other sites
Just wanted to add to my earlier post.

The brain is estimated to have one hundred billion neurons (nerve cells) that send the chemical messages that run our brains and influence our boidies. It is also estimated that there are more neurons in the average human brain than there are stars in the universe. And that some of these neurons make more than ten thousand individual connections with other neurons. So, please explain to me how one can easily estimate when exactly an individual will forget what they have been taught and then add in the fact that individual motivation on the subject under study has a lot to do with learning also.

It is based on the forgetting curve by Ebbinghaus and experimental results. Spaced repetition really has nothing to do with reverse engineering the brain from the neurons on up.

I've used spaced repetition to become much more efficient at studying Korean. There are thousands of people using spaced repetition computer flashcard programs for learning languages now who are out pacing others by miles.

So you think that personal motivation has nothing to do with how efficient someone learns? If so, I disagree, and that comes from personal experience not "experimental results."

Share this post


Link to post
Share on other sites
So you think that personal motivation has nothing to do with how efficient someone learns? If so, I disagree, and that comes from personal experience not "experimental results."

I think personal motivation has a lot to do with learning. How efficient they learn, though, is different. That is based on learning practices. Without sufficient motivation people rarely implement the most efficient learning practices since they take a lot of mental effort. Using spaced repetition software is an excellent tool for people with the motivation to consistently make daily reviews of something they want to integrate into their long term memory more efficiently than reviewing based on guesses or chance.

Share this post


Link to post
Share on other sites
So you think that personal motivation has nothing to do with how efficient someone learns? If so, I disagree, and that comes from personal experience not "experimental results."

I think personal motivation has a lot to do with learning. How efficient they learn, though, is different. That is based on learning practices. Without sufficient motivation people rarely implement the most efficient learning practices since they take a lot of mental effort. Using spaced repetition software is an excellent tool for people with the motivation to consistently make daily reviews of something they want to integrate into their long term memory more efficiently than reviewing based on guesses or chance.

Well, when I was a kid we called that studying one's material and it is no surprise that it works.

Share this post


Link to post
Share on other sites
Well, when I was a kid we called that studying one's material and it is no surprise that it works.

You used Spaced Repetition Software as a kid?

Share this post


Link to post
Share on other sites
Well, when I was a kid we called that studying one's material and it is no surprise that it works.

You used Spaced Repetition Software as a kid?

No, but what he mentions (to a certain degree) is nothing new. In general one must study in a continual manner if they are going to retain certain information.

For example, learning math as a young child. Did you attempt your math tables just once, or did you study your flash cards multiple times until you could recite them in your sleep? Well, I studied them until I knew them almost without thought and when the other 3rd graders were taking multiplication test on 3s and 4s I was on 11s and 12s.

Another example is from a young recruit that I had under me while in boot-camp. Marine recruits have to learn many different subjects and be able to recite most of this information when asked, recruits probably learn more in 3 months than most learn in a year. This specific recruit was having difficulty learning his 11 general orders along with other information. I told this recruit that if he could not learn his "knowledge" that the Drill Instructors could send him back in the training schedule and cause him to not graduate with our platoon or not at all. I then set up a schedule for him to study all night, the next day he had learned all 11 general orders along with the other information. He could recite his general orders in order, backwards or any other way demanded that he was asked. He also took the first phase test which he had just failed a couple days prior and passed with one of the highest scores. As a matter of fact my platoon (I was the guide who is the leader of a platoon of recruits) passed with the highest academic scores during my 3 months on the Island and as the guide I am the one that had to set up the learning time.

With that said, I have gotten more from Ayn Rand's writings on epistemology and Harry Binswanger's courses on psycho-epistemology than I have most other sources.

Share this post


Link to post
Share on other sites

After reading the article, I didn't get a sense of what his real insight was. Yes, they discovered that the best time to reinforce memory is at the point just before "forgetting", but I see no indication of that point having been figured out.

Or maybe this is the answer. I see this statement: "Then I found a mnemonic to enter in SuperMemo: clear/clever. Now I never misuse them." He found a mnemonic, but I guess he isn't going to say what it is.

He also says:

"The process of increasing the size of my databases gradually progressed at the cost of knowledge retention."

Okay, so what is the solution?

Also, what about this:

The Bjorks were not the first psychologists to make this distinction, but they and a series of collaborators used a broad range of experimental data to show how these laws of memory wreak havoc on students and teachers. One of the problems is that the amount of storage strength you gain from practice is inversely correlated with the current retrieval strength. In other words, the harder you have to work to get the right answer, the more the answer is sealed in memory.

Precisely those things that seem to signal we're learning well — easy performance on drills, fluency during a lesson, even the subjective feeling that we know something — are misleading when it comes to predicting whether we will remember it in the future.

This goes against everything that has worked for me, unless there is something I'm missing here.

I find this highly interesting, but there seems to be quite a bit missing here.

Brute force memorization is only one aspect of learning. Logical integration of the information you process helps with memory quite a bit. Context, hierarchy and tying things down to reality are all vital and reinforce ones knowledge, as Objectivists are well aware. I don't know if the work of Wozniak takes those sort of things into account at all.

I'll probably try some of this software out to see where it take me!

Share this post


Link to post
Share on other sites
After reading the article, I didn't get a sense of what his real insight was. Yes, they discovered that the best time to reinforce memory is at the point just before "forgetting", but I see no indication of that point having been figured out.

The point is based on a prediction (based on large samples of people) on how you rated the card.

Brute force memorization is only one aspect of learning. Logical integration of the information you process helps with memory quite a bit. Context, hierarchy and tying things down to reality are all vital and reinforce ones knowledge, as Objectivists are well aware. I don't know if the work of Wozniak takes those sort of things into account at all.

It has nothing to do with brute force memorizing. Go to Kanji.koohii.com to see thousands of people who memorize the Joyou Kanji with imaginative memory, but review it with an SRS system.

Share this post


Link to post
Share on other sites
No, but what he mentions (to a certain degree) is nothing new. In general one must study in a continual manner if they are going to retain certain information.

That is not the unique benefit of an SRS. He did not invent "review" he refined review to become efficient. You should read the article more carefully or watch this screencast.

For example, learning math as a young child. Did you attempt your math tables just once, or did you study your flash cards multiple times until you could recite them in your sleep? Well, I studied them until I knew them almost without thought and when the other 3rd graders were taking multiplication test on 3s and 4s I was on 11s and 12s.
That was inefficient. You could have reviewed less (the optimal amount, reviewing only at a time that is just before you forget them) and focussed on learning other things.

Share this post


Link to post
Share on other sites
No, but what he mentions (to a certain degree) is nothing new. In general one must study in a continual manner if they are going to retain certain information.

That is not the unique benefit of an SRS. He did not invent "review" he refined review to become efficient. You should read the article more carefully or watch this screencast.

For example, learning math as a young child. Did you attempt your math tables just once, or did you study your flash cards multiple times until you could recite them in your sleep? Well, I studied them until I knew them almost without thought and when the other 3rd graders were taking multiplication test on 3s and 4s I was on 11s and 12s.
That was inefficient. You could have reviewed less (the optimal amount, reviewing only at a time that is just before you forget them) and focussed on learning other things.

How do you know it was inefficient? The last I checked one cannot get more efficient than going through one's time tables one time and then passing them, every time.

I offer that you might get more out of Edwin Locke's guide Study Methods and Motivation: A Practical Guide to Effective Study.

Share this post


Link to post
Share on other sites
No, but what he mentions (to a certain degree) is nothing new. In general one must study in a continual manner if they are going to retain certain information.

That is not the unique benefit of an SRS. He did not invent "review" he refined review to become efficient. You should read the article more carefully or watch this screencast.

I have read and understood the article 3 times before writing my first post. I have now watched the video you linked to and I am still not convinced, nor impressed, that learning has been made more efficient. I could give you hundreds of real world examples of people learning something and retaining that information the first time. But if a person agrees with the premise from the video you linked to then I should have forgotten almost everything about my past and all the books that I have read, just once. This is obviously totally incorrect as I, and I am sure many others, have total recall of many things that I only looked at or studied one time, no matter how long ago.

Share this post


Link to post
Share on other sites

Since I posted the article, I have tried one version of the algorithm in a software called Mnemosyne. I chose that one based on some research on Supermemo. Mnemosyne is more user friendly, and it's free. It was developed in part by a memory research team that is using (anonymously) records from each instances of the software to build a database & help them in the research. This is a good trade off.

Here is how it works:

1) You work on your own to perfectly understand, and do the initial memorization of, the material you wish to study.

2) You enter flashcards yourself. For example, if you were studying French language, you would write "House", and on the "other side" you would enter "Maison". I'm using it for various computer subjects.

3) The software presents you with a selection of flashcards each day. Which flashcard it selects is based on an algorithm that tries to predicts the speed at which you are likely to forget something.

4) You rate the card from 1 to 5, with 1 and 2 meaning that you had completely forgotten the card (and that therefore the interval was way too long), 4 meaning that you were able to recall the info with some effort (which is the optimal interval, per the theory), and 5 meaning that the card was presented too early (i.e., you were able to recall too easily).

5) Over time the memory algorithm is tweaked, though I don't know exactly how.

The idea as I understand it is not that you cannot study something every day and remember it. It's that this might be a waste of time - Ray, I think this is not dissimilar from the approach you take with Progressive Exercise. So if you have time constraints (and who doesn't?), the right time to review material is (as the theory goes) JUST BEFORE you were about to forget it. Apparently, this results in stronger and stronger memory connections over time.

Of course, motivation plays a role. The material that you study hardest and are most motivated about might require more spaced time increments, if only because you might write more and better flashcards (there's an art to it, I'm getting slowly better at it).

Just as obviously, this is not going to change your nature. If you are born with a good memory, you might not see obvious benefits from that. Churchill was known to be able to recite poems he had read just once, years before. This is just the genetics at work.

At this point, I'm planning on using this software ongoing. It's working for me. It's hard to say if this is because of the way the algo works, or if it's just the method and discipline that goes into writing flash cards. I'm definitely memorizing some of the concepts better than I would otherwise.

Share this post


Link to post
Share on other sites

Joss,

I recommend you check out the program Anki. It's free and offers (I think) the easiest to use interface out there. Especially for recording audio, adding pictures, etc.

Share this post


Link to post
Share on other sites
I have read and understood the article 3 times before writing my first post. I have now watched the video you linked to and I am still not convinced, nor impressed, that learning has been made more efficient. I could give you hundreds of real world examples of people learning something and retaining that information the first time. But if a person agrees with the premise from the video you linked to then I should have forgotten almost everything about my past and all the books that I have read, just once. This is obviously totally incorrect as I, and I am sure many others, have total recall of many things that I only looked at or studied one time, no matter how long ago.

Ray, you don't understand the article and you still don't get the reason why periodic reviewing rather than constant reviewing is more efficient. It saves time. Watch the last few seconds of the video again where it shows the calendar and review rates.

Share this post


Link to post
Share on other sites
I offer that you might get more out of Edwin Locke's guide Study Methods and Motivation: A Practical Guide to Effective Study.

I have the book and it's obvious from the comment you do not understand the roll that an SRS plays. An SRS would be a reviewing adjunct to the methods applied in Locke's book. In other words, Locke would recommend that you study something and integrate it with your existing knowledge to make sure that the memory will be strong, and then you review it. The SRS would take care of how often you review, without overreviewing.

Share this post


Link to post
Share on other sites
I have now watched the video you linked to and I am still not convinced, nor impressed, that learning has been made more efficient. I could give you hundreds of real world examples of people learning something and retaining that information the first time. But if a person agrees with the premise from the video you linked to then I should have forgotten almost everything about my past and all the books that I have read, just once. This is obviously totally incorrect as I, and I am sure many others, have total recall of many things that I only looked at or studied one time, no matter how long ago.

It will remain in your short term memory if you only think about it once, unless you form a really really strong emotional bond with the knowledge (and SRS's account for that effect by having a rating system for how easily you retrieve the memory). Thinking about it once and opening the book once to formally study it are two entirely different things. Often, I can read something once and remember it forever, but that does not mean that I did not periodically think about it afterward, which is a system of reviewing.

Share this post


Link to post
Share on other sites

Duke,

You can keep spending your time figuring out when to review and I will have already learned the material. Have fun.

Share this post


Link to post
Share on other sites
Joss,

I recommend you check out the program Anki. It's free and offers (I think) the easiest to use interface out there. Especially for recording audio, adding pictures, etc.

Thanks - I have already invested enough time into Mnemosyne that it wouldn't make any sense for me to switch now. I'll take a look though.

JD

Share this post


Link to post
Share on other sites

Ray - it seems that you are a natural memory athlete. The fact that you were able to retain multiplication tables after a single instance, and are able to retain so much from your readings, makes that pretty obvious to me.

ASSUMING THE METHOD WORKS, I suspect that maybe you would see minimal benefits from the spaced repetition methods. Obviously, I don't know whether the theory is sound or the method works. I do think that the idea is worth discussing - rather than dismissing out of hand.

Also, you should consider the situation of either material that you find difficult to retain (I assume you find some material difficult to retain), or that of other people who don't have your ability / skill to memorize.

What I regret is that in this thread, there's been a lot of heat and very little light, and that's partly coming from your contributions - see in particular posts #2, 5, 7, 20. If you are not interested in the idea, or dismiss it out of hand, it would be preferable for everyone if you refrained from posting, because then it creates an atmosphere which discourages others from participating - and it wastes your time.

Share this post


Link to post
Share on other sites

Just to add to the general discussion, there are several articles on the Supermemo site that gives more insight into Wozniak's thinking, and show that he definitely agrees with the importance of motivation, grounding things in reality, building knowledge on top of knowledge, etc.

See here:

http://www.supermemo.com/articles/20rules.htm

1 - Do not learn if you do not understand

...

3 - Build upon the basics

...

13 - Refer to other memories

...

and here:

http://www.supermemo.com/articles/power.htm

1 - Nurse your hunger for knowledge - if your motivation for getting knowledge is weak you can stop reading this text. The rest of my advice will not work. Your motivation cannot be shallow (like for passing an exam, showing off at a party, impressing the boss, etc.). You have to find the clear-cut link between knowledge and the value it brings to life.

...

Share this post


Link to post
Share on other sites
Thanks - I have already invested enough time into Mnemosyne that it wouldn't make any sense for me to switch now. I'll take a look though.

Joss, I made the switch after the hard drive on my netbook failed. I've never looked back since. It is superior for studying language in my opinion, because of all the plug-ins created by the community. There are so many awesome plug-ins, like SUBS2SRS, where you can capture part of a movie and make it your question, and the subtitles become your answer. I think the reason for there being a bigger community is because Anki has been popularized by people following the AJATT method. I also love the chart outlines of statistics and stuff that it gives, the timers, and other additional niceties.

Do note that you can easily save a Mnemosyne file and import it to Anki. I would recommend you save your current cards whether or not you transfer to Anki, otherwise you could lose them like I did.

Also Joss, if you note that you have a hard time originally memorizing some things before you input them into the SRS, look up and experiment with imaginative memory or memory castles. People usually use that type of technique in conjunction with an SRS, at least for learning Chinese characters.

Share this post


Link to post
Share on other sites
Duke,

You can keep spending your time figuring out when to review and I will have already learned the material. Have fun.

Don't worry, my SRS does that for me.

Share this post


Link to post
Share on other sites