Before going into specifics, I’d like to mention that while I think attention is the keystone of the mind, I do not think it’s worth spending a majority of time thinking about. Monica Anderson has made the point before that while the human mind is both intuitive and logical, the logical portion is a tiny sliver sitting on an enormous intuitive foundation. I can make a similar claim (and perhaps Monica has made such a claim before, I don’t know) that while the mind has components that could be classified as “deliberate” or “automatic”, the automatic portions vastly outweigh the deliberate ones. (Note: I’m a determinist. “Deliberate” should not be read as non-mechanistic.) Both of these statements emphasize something I like to call the Introspective Fallacy, which is the error where we feel as though we know how our minds work because we’re conscious. (Dennet has a wonderful TED talk into to this fallacy.) And from our conscious perspective it’s damn near all deliberate and logical (more on this later). So, even though this is important stuff, I think that understanding vision, pattern recognition, muscle memory, and generative systems (imagination) overall need more… attention.

But really, what I think we need are big-pictures of the mind. And attention is an easy place to start, as it’s subjectively prominent.

Okay, so David writes:

…our ‘attention’ mechanism then searches through a . . . what would you call it? List? Sea? It seems like a tree or linked list data structure of some kind by the way your attention scans it. Like navigating a maze. Your attention searches the maze for the [original idea].

I think the structure in question here is a directed graph (“digraph”) of concepts and relationships. This is a very old theory, and those familiar with the history of AGI should immediately recognize it as a “semantic network”. Semantic nets were the first (IMO) real attempt at AGI, and they failed pretty spectacularly. But their failure, I think, should not be taken as an indicator that they don’t exist (after all, the subjective experience of searching though a “concept space” for something particular should be familiar to all humans), but rather it is an indicator that semantic nets are useless without additional structure (specifically: grounding in sensory experience).

…as [the process of attending to an idea, getting distracted, and returning to the idea] is happening, the attention and the maze or linked list function as one structure.

This confuses me. As I think of it, I imagine “attention” being a position on the semantic digraph, but that’s different from saying they’re “the same structure”. After all, what we attend to, we can analyze, while it’s impossible to analyze all our concepts simultaneously. The object of analysis must then be separate (i.e. non-equal) in some respect from the collection from which it was drawn.

The short term memory or maze that we keep all our ideas in also has a primary slot that ‘what we are thinking about right now’ is kept in.

Whoa! When I imagine the data-structure of ideas, I do not imagine it being synonymous with “short term memory”. Perhaps I misunderstand what’s being said.

I will say that it sounds like the “primary slot” is “attention”, and I could see how this is held in “short term memory”. I feel it’s important to be clear and precise about what this means, though.

[T]he apparent structure of attention/memory is separate from the analysis part of our minds. The two can act independently of one another.

While the first sentence makes some sense to me (though I don’t see how attention = memory), the second doesn’t seem possible. Isn’t analysis dependent on attention?

As a follow up, I’m curious what is meant by “analysis” from a mechanistic perspective. My intuition is that it’s very close to what I think of as “imagination”.

I think that if something really central to our mind’s operation could be proven to be part of the same ‘module’ as another part of it, that would be really significant. It would also give us a damn good starting point to put the rest of the puzzle together.

I think it’s important to keep in mind that modularity is artificial. For instance, is RAM (in a computer) a distinct module from the power supply? While almost everyone would say yes, RAM can’t function without power. Similarly, a hard-disk and the read-write mechanism can be thought of as distinct modules, but the overall storage module won’t work without both of them. Isn’t it possible to think of half of a hard-disk as a distinct module from the other half? Is your eye a distinct module from your head?

Put another way, what does modularity predict? If a person gets their cerebellum knocked out, they are still able to do complex math, so we can theorize that the cerebellum is not a part of the “math module”, but by this same logic, does it make sense to say that breathing is part of the “math module”? I think if the terms are hammered down and made concrete, the experimental science can manage the rest of the way.

Is attention really the same thing as short term memory?

It seems true that what we attend to is stored in short-term memory, but it does not make sense to say that they’re synonymous. To make the falsification explicit, imagine a person who is shown a card, then this person attends to something entirely different. Might the person remember their card when shown it again? Yes. Ergo, short-term memory must be more than attention.

Is attention separate from analysis to such a degree that they run in parallel?

It all depends on the terms, of course, but I think a big part of the answer comes from considering the alternative. What would attention and analysis running sequentially look like? Would a thing stop being attended to while analysis was going on? It seems almost paradoxical from my perspective, as it seems to me that “attention” is what we are currently “analyzing”. But again, this is a terminology issue. I think grounding the language as predictions would solve things.

Just as important: Do you know of any ideas or papers that talk about the ‘structures of the mind’ like the ones I’m talking about…

Unfortunately, I’ve found very little in the way of broad theories of what I call “mental architecture”. Most neuroscientists seem to enjoy burrowing into specific regions of the brain such that “thought” becomes a fairy tale, and most psychologists seem content to ignore the “how” of the mind, and focus entirely on the “what” of behavior.

The closest I’ve come has been Kosslyn and Koenig’s Wet Mind, though I admit that I’ve only gotten about half way though it so far. Though it’s hugely lacking in certain respects, Jeff Hawkins’ On Intelligence is pretty good, too.

As a final note, David brought up the question of consciousness, which is such a painfully undefined and scattered concept that I’ve intentionally avoided it in this post. I have my own pet theories of consciousness, but they’re not worth addressing without the firm backdrop of a strong mental architecture, I think.

If someone is interested in the truth, it’d be a blessing to them to learn exactly what falsehoods they believe. An ideal truth-seeker would love to be wrong, because it would give them the chance to improve their ideas. This is the basis of scientific thinking, but it’s not a very good description of how people actually behave. Most people (especially those with little training) react defensively to criticisms of their ideas and show a strong preference of comfortably holding familiar ideas rather than putting them to the test. Why is this?

Some of this bias can be explained by a desire for familiarity (“if democrats are wrong than my whole world-view would fall apart”) and an attachment of self-esteem to being right (“only bad people are wrong”), but I think I’ve stumbled across another, more biological, explanation.

Imagine some ancestral hominids try two kinds of fruit: one kind that is delicious, and one kind that is poisonous. Those who eat the poisonous fruit spend all night in terrible pain, but eventually recover and learn to eat the good fruit. Twenty years later, some of the same hominids return to the grove with the two fruit. It’s been a while, so maybe they’ve forgotten about the good fruit, but they certainly haven’t forgotten about the poisonous one. The reason for this is that brain weights memories surrounding unpleasant events more heavily than pleasant ones. In short: it’s more important to remember what could’ve killed you or your family than just about anything else.

This mnemonic weighting is present in rats and other animals only distantly related to humans, so it’s sensible to conclude that it developed in a very primitive animal, far before logical reasoning, language and culture. When it became possible to reason about the world, humans were stuck with their old equipment for remembering. It was still more important to remember mistakes (even if they were mistakes of reasoning), but rather than build an emotionality-neutral way to add weight to such fallacies, evolution took the shortcut of adding unpleasant feelings to logical errors. By making it annoying and embarrassing to be wrong, evolution ensured we’d learn our lesson. Many years later, when scientific thought has become important, we’re left with the reality that, by default, people tend to avoid criticism.

(Photo credit: http://www.flickr.com/photos/idmaer/86349266/ Licensed under Creative Commons Attribution-Noncommercial-Share Alike 2.0 Generic)

Continuing my search for AGI I’ve been thinking about conscious and unconscious thought. Investigating these led me to explicit and implicit memory. This actually led me to the lecture by Eric Kandel (above), which goes into good detail the mechanisms by which animals learn. It’s not complete, though. Kandel only describes learning as an emphasizing of existing synapses, rather than the formation of connections to new neurons. To make things worse, he admits that this is not well understood by the scientific community… so I’m still mostly stuck.

All this pondering has led me to some interesting ideas, though. Earlier, I had associated conscious thought with logical thought, and unconscious with intuitive. Branching out from there I began to wonder if logical thought was the product of explicit memory, and intuitive from implicit memory. It now seems clear to me that this cannot be such a clearly-defined dualism, because mice possess an explicit memory and are not competent at logic. Logic is a procedure that is perhaps only available to conscious entities, but I should not jump to the conclusion that consciousness entails logic, either, as it seems to me that a lot of the conscious (deliberate) action taken by humans is not based on theory.

Humans possess both explicit and implicit memories, and have the capability to use intuition or logic to solve problems. If I’m right, and strong artificial intelligence requires replicating the capability of non-human animals, then I’d do well to first disregard logic, then disregard explicit learning of events and facts, and focus instead on how novel neural links form in nerves.

Of course, I always try to assume I’m wrong. So in hopes I’ll figure out how, I’m off to watch the syntience (artificial intuition) videos.