Augmenting Long-term Memory

I use anki for around seven years now and it has proven to be the best learning technique for me so far.

When I went to university I still believed that understanding is everything and learning will happen en passant. I read many books, pondered long about them, and forgot everything a month later.

Five years ago I started studying again (computer science) and as anki has already been usefull in learning programming for me I used it extensively during my studies. Instead of taken notes on paper I just marked important facts while reading and made anki cards for them afterward. Sometimes these where simple atomic facts like "What is the definition of XYZ?", sometimes they demanded a longer answer "How does x yield y?". It took me about 20 to 30 minutes each day to reherse my cards (while commuting to work) and learning has never been that easy.

I soon realized that understanding a fact or each step of a proof the first time you read it is not as important as often stated. Understanding often only came with learning because anki forced me to think about these facts daily. That was much more effective than pondering hard about them once.

I made a deck for each course the day it started and archived it after the exam. But not only did this help me during the exam I also forgot far less of what I have learned afterwards. Overall I would say that learning with flashcards took only about a fifth of the time it would have taken with other learning techniques that have a stronger focus on understanding (mind-maps, active reading &c.)

If somebody is looking to streamline their Anki/Mnemosyne card creation, I created this vim plugin a while back:

https://github.com/tbabej/knowledge

It takes a plain text note file and converts it to a set of cards. Syncs the updates if the notes change, and also supports things like images.

I don't have as much time as I used to for open source development, and could certainly use some help (especially on the Anki side, as I am a Mnemosyne user now).

Note that this allows you to not only have a set of flash cards, but also a nice collection of notes grouped by topic, if you want to have an overall look at the section of your domain knowledge (say, linear algebra) in context. This is a bit hard to do with a set of cards already entered in a flash-card application.

Fantastic article!

I want to point out that the essence of these systems, Anki and the others, is that you are teaching your brain what you value.

Your brain is an organ, a kind of gland for thought, and it uses a huge amount of energy (in the form of calories and oxygen, your brain uses something like 20%-- one fifth --of all the oxygen you breath) to do its job.

The brain is only going to spend as much energy as it has to for you to survive, because evolution made it that way. It's only going to spend its energy budget on memory of "facts" if it's convinced there's important survival value to do that.

This budget is largely automatic, but you can influence it, and these flash-card systems work by doing that.

The time and attention you put into setting this up and then being able to successfully answer the questions that pop up, exactly mimic the time, attention, and reward of hunting small game.

Think about how a cat studies anything mouse-like, and compare that to the way most students treat studying...

Anki, et. al., set up a scenario where your brain is being trained to regard memory and recall of facts as a fun and easy way to scratch that survival itch.

Your brain can actually remember things instantly and permanently and without conscious effort (in the extreme, pathological case we get phobias) but it won't bother unless it's convinced you care.

"Most students will not use such trivial devices as flash cards; it seems to be beneath their dignity. They suffer accordingly."

Richard Hamming

This article is about committing knowledge to long term memory. I've been thinking about ways to actually augment long term memory with external storage. I'm unsatisfied with current software (mind maps, OneNote/Evernote, WorkFlowly, org-mode...). They all feel like building a document. I noticed when I come back to my notes in those formats it takes too much time to retrieve information and to add new information without re-structuring existing information. The existing solutions are not optimized for quickly putting down some facts/ideas/goals for later retrieval, without worrying about layout and structure.

To efficiently augment memory, the software (especially UI) should mimic the way we think. Each captured thought should be in some way related to one or more previous thoughts and stored in the correct context. Inputted notes should be fairly short and they should be organized in graph structure. Edges should define relations, for example: contains, depends on, implies, follows.

When retrieving a thought it should be visualized in context with other related thoughts. This graph should have same layout each time it's retrieved, for better visual navigation. If the system limits the visualization of graph up to 2nd neighbors, there should always be enough space on 2D plane to expand the graph with new thoughts. Manual layout should be discouraged because it wastes too much time.

The basic storage/retrieval model could also be expanded with additional processing to further offload brain activities: logical processing (if an assumption proves to be incorrect, all dependent thoughts should be flagged as incorrect/uncertain), goal prioritization, future event reminders.

The hard problem is entry method. Because such system should always be accessible, the candidate devices are smartphone and smart watch. So far all input methods except keyboard are too slow and error prone. Most input methods also obscure much of screen space and require visual feedback to verify that text is correct (swiping and auto-correct). I'm researching gesture-based and chording virtual keyboards, but there's nothing suitably fast.

This is on my side-project back log, but hopefully I'll find something close enough I can use instead of building from scratch. Any suggestions?

Great article! I'm a long time Anki user, and while our use patterns for Anki are a bit different, I really identify with almost everything in that article.

I highly recommend you learn and use Anki, especially if you're actively studying something (e.g., for a degree). I'm using Anki to (partially re-)learn mathematics, and it is gratifying to know that, with a reasonable amount of certainty, whatever I learn I'll actually remember for the long term (unlike when I did my CS degree).

The beauty of human memory is not what we remember but what we forget.

What is the current state of collaborating on Anki decks? I see there is the CrowdAnki plug-in – this exports decks to a JSON file, which can be version controlled and uploaded to Github. But how well does this work in practice? I feel it would be nice to have a more integrated way of collaborating on decks.

This is great. Anki looks amazing. The reason I don't use Anki, though, is that I feel like the things I wish I could remember are rarely things I would have had enough foresight to make a flashcard for. In other words, I rarely feel like I have a problem remembering things that I would have been wise enough to make a reminder about. The things I wish I could remember are things in books that I didn't even pay much attention to on a first reading. It's things that I wish I could see all the hidden connections to once I realize some new insight.

There are things I know I want to remember and there are things I realize I would have wanted to remember more clearly only after I have learned some fundamentally new way of viewing the world. The former Anki solves but they're rarely things I feel a problem with, and the latter is by definition things I wouldn't have paid enough attention to to write down in Anki.

I augment my long-term memory by truly augmenting my brain via blog posts on ryanmercer.com that I tag thoroughly, Evernote entries that are filed broadly and tagged thoroughly and by forwarding remotely useful emails to myself in Gmail with the search strings I'd likely use to find them in the future.

Obviously this requires connectivity but I'd rather offload the stuff to a server somewhere than carry EVERYTHING in my mind.

Does anyone know if there is an alternative to Anki with a better web experience? My biggest hurdle in using Anki is working on computers on which I am unable to install software and the lack of quality support for the web version.

Anyone have a link for where to obtain the desktop and mobile clients? I couldn't find links in the page, and Google play store has dozens of things named Anki.

This is fascinating. I'm starting a new job soon with a new platform to learn and this is something I'm going to incorporate.

windows client installer is pretty large at 29MB and scores 3 detections out of 66 on virustotal.

Remembering is not as important as training. As recent HN submission put it, you can't tell people anything. You can, however, learn languages with anki-style cards, if that is what you want.

My experience with Anki and similar programs has always been that my ability to recall the learned information is heavily dependend on context. Words or facts I can perfectly recall in Anki often didn't come to me in real-life situations.

It would be interesting to automate the “ankyfication” process, i.e. given a text extract a set of questions and answers , then it can be used either for personal learning or as a feed to a memory machine