(2018-07-31) Nielsen Augmenting Long Term Memory

Michael Nielsen: Augmenting Long-term Memory. Alexander Luria went on to study Shereshevsky's memory for the next 30 years. In a book summing up his research Alexander Luria, “The Mind of a Mnemonist”, Harvard University Press (1968)., Luria reported that: [I]t appeared that there was no limit either to the capacity of S.'s memory or to the durability of the traces he retained.*

Given how central memory is to our thinking, it's natural to ask whether computers can be used as tools to help improve our memory.

Vannevar Bush, As We May Think, The Atlantic (1945). for a mechanical memory extender, the memex... It is an enlarged intimate supplement to his memory

The memex vision inspired many later computer pioneers, including Douglas Engelbart's ideas about the augmentation of human intelligence, Ted Nelson's ideas about hypertext, and, indirectly, Tim Berners-Lee's conception of the world wide web

From the memex to the web to wikis to org-mode to Project Xanadu to attempts to make a map of every thought a person thinks (HowToMakeACompleteMapOfEveryThoughtYouThink): the augmentation of memory has been an extremely generative vision for computing.

In this essay we investigate personal memory systems, that is, systems designed to improve the long-term memory of a single person

In the first part of the essay I describe my personal experience using such a system, named Anki.

The second part of the essay discusses personal memory systems in general.

make a case that memory is central to problem solving and creativity. Also in this second part, we'll discuss the role of cognitive science in building personal memory systems and, more generally, in building systems to augment human cognition.

You can reasonably think of the essay as a how-to guide aimed at helping develop virtuoso skills with personal memory systems.

Part I: How to remember almost anything: the Anki system

At first glance, Anki seems nothing more than a computerized flashcard program

for my personal set of Anki cards the average interval between reviews is currently 1.2 years, and rising

I therefore have two rules of thumb. First, if memorizing a fact seems worth 10 minutes of my time in the future, then I do it

Second, and superseding the first, if a fact seems striking then into Anki it goes, regardless of whether it seems worth 10 minutes of my future time or not.

Anki makes memory a choice.

I use Anki in all parts of my life. Professionally, I use it to learn from papers and books; to learn from talks and conferences; to help recall interesting things learned in conversation; and to remember key observations made while doing my everyday work. Personally, I use it to remember all kinds of facts relevant to my family and social life; about my city and travel; and about my hobbies.

I've used Anki to create a little over 10,000 cards over about 2 and a half years of regular use

When I'm keeping up with my card review, it takes about 15 to 20 minutes per day.

At a practical level, I use the desktop Anki client for entering new cards, and the mobile client

I had trouble getting started with Anki

What made Anki finally “take” for me, turning it into a habit, was a project I took on as a joke. I'd been frustrated for years at never really learning the Unix command line. I'd only ever learned the most basic commands. Learning the command line is a superpower for people who program, so it seemed highly desirable to know well. So, for fun, I wondered if it might be possible to use Anki to essentially completely memorize a (short) book about the Unix command line.

got me to experiment with different ways of posing questions.

Using Anki to thoroughly read a research paper in an unfamiliar field

After the match where AlphaGo beat Lee Sedol, one of the strongest human Go players in history, I suggested to Quanta Magazine that I write an article about the system

While I was excited, writing such an article was going to be difficult. It was going to require a deeper understanding of the technical details of AlphaGo than a typical journalistic article.

I began with the AlphaGo paper itself. I began reading it quickly, almost skimming

Here's a few examples of the kind of question I entered into Anki at this stage: “What's the size of a Go board?”; “Who plays first in Go?”; “How many human game positions did AlphaGo learn from?”

I made several rapid passes over the paper in this way, each time getting deeper and deeper

After five or six such passes over the paper, I went back and attempted a thorough read. This time the purpose was to understand AlphaGo in detail.

Many of the questions I was putting into Anki were high level, sometimes on the verge of original research directions.

This entire process took a few days of my time, spread over a few weeks.

I find Anki works much better when used in service to some personal creative project

goals which, for me, are intellectually appealing, but which I'm not emotionally invested in. I've tried this a bunch of times. It tends to generate cold and lifeless Anki questions, questions which I find hard to connect to upon later review, and where it's difficult to really, deeply internalize the answers

Using Anki to do shallow reads of papers

This doesn't mean reading every word in the paper. Rather, I'll add to Anki questions about the core claims, core questions, and core ideas of the paper.

Many papers contain wrong or misleading statements, and if you commit such items to memory, you're actively making yourself stupider.

Another useful pattern while reading papers is Ankifying figures.

I have an Anki question which simply says: “Visualize the graph Jones 2011 made

I count myself as successful if my mental image is roughly along those lines

Really good resources are worth investing time in. But most papers don't fit this pattern, and you quickly saturate. If you feel you could easily find something more rewarding to read, switch over. It's worth deliberately practicing such switches, to avoid building a counter-productive habit of completionism in your reading.

Syntopic reading using Anki

“read” the entire research literature of some field or subfield.

You might suppose the foundation would be a shallow read of a large number of papers. In fact, to really grok an unfamiliar field, you need to engage deeply with key papers – papers like the AlphaGo paper.

So, to get a picture of an entire field, I usually begin with a truly important paper, ideally a paper establishing a result that got me interested in the field in the first place. I do a thorough read of that paper, along the lines of what I described for AlphaGo. Later, I do thorough reads of other key papers in the field – ideally, I read the best 5-10 papers in the field. But, interspersed, I also do shallower reads of a much larger number of less important (though still good) papers.

this is a form of what Mortimer Adler and Charles van Doren dubbed syntopic reading* In their marvelous “How to Read a Book”*

I start to identify open problems, questions that I'd personally like answered, but which don't yet seem to have been answered. I identify tricks, observations that seem pregnant with possibility, but whose import I don't yet know. And, sometimes, I identify what seem to me to be field-wide blind spots

In this way, Anki is a medium supporting my creative research

with a field I already know well, my curiosity and my model of the field are often already so strong that it's easy to integrate new facts. I still find Anki useful, but it's definitely most useful in new areas.

Anki helps give me confidence that I can simply decide I'm going to read deeply into a new field, and retain and make sense of much of what I learn

One surprising consequence of reading in this way is how much more enjoyable it becomes. I've always enjoyed reading, but starting out in a challenging new field was sometimes a real slog, and I was often bedeviled by doubts that I would ever really get into the field. That doubt, in turn, made it less likely that I would succeed.

More patterns of Anki use

Make most Anki questions and answers as atomic as possible

Breaking this question into more atomic pieces turned a question I routinely got wrong into two questions I routinely got right

I'm not sure what's responsible for this effect. I suspect it's partly about focus. When I made mistakes with the combined question, I was often a little fuzzy about where exactly my mistake was

One benefit of using Anki in this way is that you begin to habitually break things down into atomic questions. This sharply crystallizes the distinct things you've learned

one real benefit is that later I often find those atomic ideas can be put together in ways I didn't initially anticipate

Anki use is best thought of as a virtuoso skill, to be developed

Anki isn't just a tool for memorizing simple facts. It's a tool for understanding almost anything.

many of the observations I've made (and will make, below) about how to use Anki are really about what it means to understand something. Break things up into atomic facts. Build rich hierarchies of interconnections and integrative questions. Don't put in orphan questions. Patterns for how to engage with reading material. Patterns (and anti-patterns) for question types. Patterns for the kinds of things you'd like to memorize. Anki skills concretely instantiate your theory of how you understand; developing those skills will help you understand better

Use one big deck

it's good to collide very different types of questions

Avoid orphan questions

too disconnected from my other interests, and I will have lost the context that made me interested.

It's not bad to have a few orphan questions in Anki

Don't share decks

Construct your own decks:

The act of constructing an Anki card is itself nearly always a form of elaborative encoding

With that said, there are some valuable deck-sharing practices. For instance, there are communities of medical students who find value in sharing

But for deeper kinds of understanding, I've not yet found good ways of using shared decks.

Cultivate strategies for elaborative encoding / forming rich associations (associative):

What about memory palaces and similar techniques?

They seem less well developed for more abstract concepts

it may be worth further investigating some of the techniques used by practitioners to form rich associations. As Foer says, quoting a memory expert, there is great value in learning to “think in more memorable ways”.

95% of Anki's value comes from 5% of the features

My cards are always one of two types: the majority are simple question and answer; a substantial minority are what's called a cloze: a kind of fill-in-the-blanks test.

Anki offers something like a 20-fold improvement over (say) ordinary flashcards

The challenges of using Anki to store facts about friends and family:

Using a memory aid feels somehow ungenuine, at least to me.

Procedural versus declarative memory:

to really internalize a process, it's not enough just to review Anki cards. You need to carry out the process, in context. And you need to solve real problems with it.

Even better would be a memory system that integrates into my actual working environment.

Getting past “names don't matter”:

they're the foundation that allows you to build up a network of knowledge.

What do you do when you get behind?

Using Anki for APIs, books, videos, seminars, conversations, the web, events, and places:

For instance, for seminars I try to find at least three high-quality questions to Ankify. For extended conversations, at least one high-quality question to Ankify. I've found that setting quotas helps me pay more attention

I tend to Ankify in real time as I read papers and books

One caution is with books: reading an entire book is a big commitment, and adding Anki questions regularly can slow you down a lot.

Something I haven't yet figured out is how to integrate Anki with note taking for my creative projects. I can't replace note taking with Anki – it's too slow, and for many things a poor use of my long-term memory

Part of the problem is that I don't have a very good system for note taking, period!

Avoid the yes/no pattern

Aren't external memory aids enough?

for creative work and for problem-solving there is something special about having an internalized understanding. It enables speed in associative thought, an ability to rapidly try out many combinations of ideas, and to intuit patterns, in ways not possible if you need to keep laboriously looking up information.

If personal memory systems are so great, why aren't they more widely used?

Systems such as Anki are challenging to use well, and easy to use poorly.

Part II: Personal Memory Systems More Broadly

In the second, briefer, part of this essay we'll consider two broader questions about personal memory systems: how important is memory as a cognitive skill; and what is the role of cognitive science in building personal memory systems?

How important is long-term memory, anyway?

I won't defend bad classroom teaching, or the way organic chemistry is often taught. But it's a mistake to underestimate the importance of memory. I used to believe such tropes about the low importance of memory. But I now believe memory is at the foundation of our cognition

My somewhat pious belief was that if people focused more on remembering the basics, and worried less about the “difficult” high-level issues, they'd find the high-level issues took care of themselves.

I had no idea at all how strongly it applied to me. Using Anki to read papers in new fields disabused me of this illusion. I found it almost unsettling how much easier Anki made learning such subjects

One striking line of work was done (separately) by the researchers Adriaan de Groot and Herbert Simon, studying how people acquire expertise, focusing particularly on chess

Now, the concept of chunks used by Simon in his study of chess players actually came from a famous 1956 paper by George Miller, “The Magical Number Seven, Plus or Minus Two”

Exactly what Miller meant by chunks he left somewhat vague

in Miller's account the chunk was effectively the basic unit of working memory.

Distributed practice

In this section we briefly look at one of the key underlying ideas from cognitive science, known as distributed practice.

memories decay

exponentially with time

probability will decay exponentially after the re-test, but the rate of decay will be slower than it was initially. In fact, subsequent re-tests will slow the decay still more, a gradually flattening out of the decay curve

On the role of cognitive science in the design of systems to augment cognition

While scientists have done a tremendous number of studies of distributed practice, many fundamental questions about distributed practice remain poorly understood.

we have enough understanding of memory to conclude that a system like Anki should help a lot. But many of the choices needed in the design of such a system must be made in an ad hoc way, guided by intuition and unconfirmed hypotheses.

The experiments in the scientific literature do not yet justify those design choices

As a consequence, system designers must look elsewhere, to informal experiments and theories. Anki, for example, uses a spacing algorithm developed by Piotr Wozniak on the basis of personal experimentation

In practice, what you want is bold, imaginative design, exploring many ideas, but inspired and informed (and not too constrained) by what is known scientifically

The human-computer interaction (HCI) community has tried to achieve it in the systems they build, not just for memory, but for augmenting human cognition in general. But I don't think it's worked so well. It seems to me that they've given up a lot of boldness and imagination and aspiration in their design

It's telling that publishing conventional static papers (pdf, not even interactive JavaScript and HTML) is still so central to the field. . At the same time, they're not doing full-fledged cognitive science either.

This suggests to me the need for a separate field of human augmentation (Augmenting Human Intellect). That field will take input from cognitive science. But it will fundamentally be a design science, oriented toward bold, imaginative design, and building systems from prototype to large-scale deployment.