Graham Glass, etc.

EDU 2.0: Biggest month so far

We added more than 6,000 new users to EDU 2.0 in the last month, which make it our biggest month so far. It's particularly encouraging because most schools are currently on vacation and so June and July are typically very slow for us.

Here's a chart of our growth, which shows us breaking the 56,000 mark a week ago:

EDU 2.0: three more color schemes

We just added three more color schemes to EDU 2.0, bringing the total number of "skins" to five. Here are some screenshots (click on an image to see the full-size version).

"Green":

"Magenta":

"Grey":

Ellipse

I'm excited! The new Imogen Heap album "Ellipse" is due out on Aug 25th.

Check out the trailer:

Free Will, Part 6

In the last part of this series I said a deterministic brain would generate thoughts based on things like its previous thoughts, its environment, its genetics, and its memories. Practically speaking, though, it would remain unpredictable, even to itself, due to things like chaos theory (which apply even in a completely deterministic Universe).

Since it's the inner workings of a deterministic brain that make decisions rather than an overwhelming force from the outside world, such a brain can be said to have Free Will.

Some people, however, prefer the idea of a mind that can have thoughts without a preceding chain of cause and effect.

I'm assuming for the sake of discussion that we're not talking about a brain that simply has some quantum non-determinism thrown in, because the effect of this on the brain's operation would probably not even approach the existing effects of chaos theory. So we're really talking about a brain that can generate complex thoughts without prior cause.

An extreme case of this would be a person who spontaneously comes up with the theory of relativity without any background in physics. Or a person who sketches out a design for a digital intelligence without any background in A.I. or cognitive science. To the best of my knowledge, no such event has ever been recorded.

Getting to the heart of the matter, my impression is that "Free Will purists" don't believe that a deterministic brain can be creative. While they accept that a deterministic brain can react in simple ways to events, they believe that something special and outside the laws of physics must be responsible for creativity.

So in the next part of this series I'll discuss the essence creativity and how a deterministic brain can be creative.

Free Will, Part 5

In the last part of this series I stated that if the Universe is deterministic and thoughts are structures within the Universe, then it follows that our thoughts are a result of a chain of physical cause and effect.

What are the consequences of this?

It means that your thoughts are the result of things in the past like:

- your previous thoughts
- your beliefs
- your childhood
- your genetics, since they influence your brain structure
- your environment
- the things that are going on around you
- how you exercise, since that influences your body chemistry
- what you eat

What impact does this have on the concept of free will? Well, it depends on which definition of free will you prefer. Here are some definitions from the web:

• the power of making free choices unconstrained by external agencies
• the doctrine that we are able to choose our actions without being caused to do so by external forces
• the partial freedom of the agent, in acts of conscious choice, from the determining compulsion of heredity, environment and circumstance

The overriding theme of free will is being able to make choices that are not determined purely by external forces. As I mentioned above, the thoughts of a deterministic brain are heavily influenced by its previous thoughts. Indeed, the physical structure of a brain changes continuously as a result of its thought processes. So a large part of what goes into a thought generated at time T is internal to the brain at time T. So assuming that the word "external" in the definitions of free will I listed above means "external at the time of making a choice", then a deterministic brain has free will.

I get the impression that "free will purists" don't feel that we have true free will unless we can have thoughts that don't require a physical chain of events to cause them. This stance requires the word "external" to be interpreted in an extreme sense to mean "external at any time during the existence of the brain".

While I understand the initial appeal of this desire, I think it's worth a closer examination. I'll discuss the consequences of thoughts-without-physical-cause in the next part of this series.

Free Will, Part 4

In the last part of this series I pointed out that even if the Universe is deterministic, it cannot be predicted to any degree of accuracy. The future is therefore mostly opaque to us, regardless of how much insight or computer power we have available to us.

Now let's consider the nature of brains and indeed all other things that are part of the Universe.

According to the Big Bang theory, the initial Universe was an incredibly hot soup of particles. This soup cooled down enough in the first few minutes to allow protons and neutrons to form via a process of fusion. This process stopped after about 17 minutes and then it took another 250,000 years before atoms started to form.

The force of gravity gradually pulled the atoms into clumps which formed quasars after about 150 million years. A billion years later, stars also started to form. Our own solar system was formed about 8 billion years ago.

There is nothing about this story of the Universe that is incompatible with determinism. The state of each particle, neutron, atom, molecule, rock, planet and star could be the result of an extremely long and complex chain of events.

As time marched on, more complex structures evolved such as replicating molecules, RNA, DNA, cells, plants, insects, and animals. Even thoughts are structures. And just like atoms and molecules, each could be the result of a chain of events. Complex structures have a much richer set of behaviors but they're still ultimately made of the same stuff.

If the Universe is deterministic and thoughts are structures within the Universe, then it follows that our thoughts are a result of a chain of cause and effect. In the next part of this series I'll consider the consequences of this observation.

School Districts

Now that edu 2.0 has nice support for individual schools, we decided to add support for school districts. A school district is essentially a network of schools, and we're going to add a ton of related features such as resource sharing within a district, district-wide groups, messaging, etc.

Here are a couple of screen shots that show the early version of support for districts. Note the new 'District' tab next to the 'School' tab.

Here's the "Wizard's" district:

Two schools have joined the district so far:

Free Will, Part 3

In this part of the series I examine whether determinism implies predictability.

In simple deterministic systems, like a ball moving through a vacuum, you could measure the ball with a laser and calculate its future position. You would have to be careful, however, to take the force of the laser measuring device into account! This illustrates a more general problem; if you want to predict a system, the act of observation changes the system which in turn must be factored into your predictions.

In fact, it's even worse that this. Heisenberg's uncertainty principle states that it's theoretically impossible to know the exact position and velocity of a particle at the same time. The more precisely you know one of these values, the less precisely you can know the other.

But wait, it gets even worse! Chaos Theory shows that tiny perturbations in a large dynamic system can trigger a cascade of events that leads to a large system-wide change. The Butterfly Effect movie did a good job of illustrating this. One extreme example of this would be: how different would the world have been if a random cosmic ray had fried the navigation system in Hitler's plane and it had crashed before he started World War II? Even without Heisenberg's uncertainly principle, chaos theory makes large dynamic systems virtually impossible to predict.

So even if the Universe is deterministic, the combination of Heisenberg's uncertainty principle and chaos theory make it theoretically impossible to predict exactly how things will unfold. I will use this fact when analyzing the rest of the question related to Free Will.

Free Will, Part 2

In part 2 of this series, I address the question "Is the Universe Deterministic?".

If everything in the Universe interacted according to classical Newtonian mechanics, then the answer would be "Yes". The Universe would basically act like a vast pinball machine, and the positions and velocities of every particle at time T+1 would be based on the positions and velocities of the particles at time T.

However, quantum mechanics introduces some randomness into the system whereby positions and velocities can only be expressed as probabilities. A straightforward interpretation of quantum mechanics implies that the Universe unfolds probabilistically and is not a pinball machine, making the answer to our question "No".

Some scientists believe, however, that even quantum mechanics is deterministic and there are non-local hidden variables behinds the scenes that would make the Universe deterministic again.

So the answer to "Is the Universe deterministic?" is currently "Maybe".

In the next part of this series I'll examine whether the answer to this question even matters when discussing Free Will.

Free Will, Part 1

A common philosophical question is "Do we have free will?" This question is often asked in the context of discussions about determinism, which is the concept that every event is a direct consequence of events that preceded it.

The line of questioning is usually "If the Universe is deterministic and we know the initial conditions at the time of the Big Bang, then everything after that is mathematically predetermined. If this is the case, then there's nothing we can do to change the future and our concept of free will is just an illusion".

The rest of this series will systematically analyze this line of reasoning, starting with the question "Is the Universe deterministic?"

New Styles for EDU 2.0

We've started to release some new "professional" styles for "edu2.0 for school" in preparation for the launch of "edu2.0 for business". The first style is "Turquoise", and here are a couple of screen shots. We'll be releasing three more styles over the next week or so.

School home page:

User home page:

Emotions

Many people believe that emotions are mysterious things that will prevent us from ever creating a Digital Intelligence. I believe that emotions in fact easier to understand that most other facets of a mind and will be a basic part of any intelligence that we build.

Here's why.

We feel emotions under certain conditions. For the purposes of this posting I'll focus on three emotions - surprise, frustration, and success. For example, we experience surprise when something unusual happens that we did not predict. Similarly, we experience frustration when we're having difficultly achieving a goal, and experience success when we accomplish a goal.

The question is: how do emotions help us?

Let's take the case of surprise. When something unexpected occurs, we don't know exactly why it happened. So in order to improve our chances of figuring out the underlying cause, it's important that we remember all the things that were happening around the time of the surprise so our mind can use this information when deducing the possible reasons. Assuming that our mind is constantly generating predictions and comparing them against observed events, it's easy to imagine our brain circuitry generating a localized "surprise" signal when a prediction doesn't match. This signal propagates quickly from the point of detection to areas throughout the brain. When an area receives a "surprise" signal, it will usually remember events especially well during that window of time to help the brain deduce the reason for the surprise. In addition, some areas of the brain respond to the signal by using face/body language to convey our state of surprise to the outside world. This has the benefit of harnessing others to also try and figure out the reason for the surprise. For example, if a Mother notices that her child is experiencing surprise, she will often try and figure out what the child is surprised by and then teach the child the reason for the particular event.

So in the case of surprise, the reason for the emotion is to harness the rest of the brain to try and figure out the reason for the surprise, as well as other people in the vicinity who might be able to help.

Now let's take the case of frustration. Once again, assuming that our minds have circuitry to detect the progress of a particular goal, it's easy to imagine the circuitry generating a "frustration" signal when a goal is taking too long. Like surprise, frustration propagates from the point of detection to areas throughout the brain. When an area of the brain receives a "frustration" signal, it might be able to come to the aid of the frustrated area and help solve the goal. In addition, if someone nearby notices that a person is frustrated, then will often come to help that person achieve the goal. For example, if a Father notices that his child is frustrated because he can't put his shoes on, he will often help the child to achieve that goal.

Once again, in the case of frustration the reason for the emotion is to harness the rest of the brain and any surrounding people to help achieve the goal that is taking too long.

Finally, let's take the case of success. When you achieve a goal, especially a difficult one, you experience success. The "success" signal is generated when the brain circuitry satisfies a goal. The signal propagates from the point of detection to many areas of the brain and rewards the areas that helped to achieve the goal. In addition, the expression of joy on your face causes a similar signal to be triggered in other people (if any) who were helping you to accomplish the goal, thereby also rewarding them.

Notice that in all three cases the emotional "signal" was initially triggered by brain circuitry that is specially designed to detect a particular internal state. The signal propagates throughout the brain and then continues to propagate to nearby people via face/body language. The signal causes areas within the brain to address the cause of the signal as well as engaging other minds in the vicinity.

I've left out a fair amount of detail in the preceding paragraphs, but I hope that I've made one point clear; emotions are vital for organizing the activity of a brain (and surrounding brains) to deal with a particular event. Indeed, it's hard to imagine a mind that doesn't use this approach.

Now that I've outlined the high-level reason for emotions, it's worthwhile pointing out that most people ultimately agree that they're useful. The thing that people have a particularly hard time with is why a particular emotion feels the way it does. In other words, why does frustration feel bad and why does success feel good? And similarly, even if a digital intelligence could generate a signal that represents frustration, how could it ever feel frustrated?

The last question is equivalent to the question "could a digital intelligence ever experience the color red?", and it sometimes called "the hard problem". I like to separate the practical role that emotion plays (which "mind engineers" could implement as described above) from more philosophical "hard problem" because it's useful to break problems down into smaller pieces and then solve each one.

I will address (and try to solve) "the hard problem" later in this series.

My LASIK Experience

A few weeks ago I decided to research LASIK eye surgery. I was recommended the Pacific Vision Institute by my ophthalmologist because they have a great reputation and use the latest technology. I had a free consultation with them and decided to go ahead and book the operation.

For the next 5 days I wore my glasses instead of contact lenses, had a screening to make sure my eyes were good candidates, and kept my eyes well lubricated. Everything went well and so I had the LASIK surgery today at noon.

Here's what it was like.

First of all, they give you a mild sedative to calm your nerves. I was fairly relaxed already, but it definitely helped. Then they dab your eyes with an anesthetic and give it a few minutes to work. Then they move you to the first machine, which uses a laser to cut a circular flap in your cornea that will be pulled back to let the second machine do the main work.

The doctor (Dr. Faktorovich) has a great bedside manner and explained everything that she was about to do. She put a ring into my eye that prevented it from closing and then cleaned and lubricated the eye. Finally, she lowered the machine over the eye until it was bound to the eye (perhaps via some kind of connector; I couldn't see that much). They told me to look at the light and then the laser started to cut the flap.

At that point, your eye goes dark and you can see lots of pretty colored dots. I thought they might be caused by the laser, but apparently it's due to the pressure of the machine against the eye. There was no pain or discomfort during this process, and it took about 10-20 seconds per eye. At this point, each eye had a corneal flap that was ready to be lifted.

A nurse then guided me from the first machine to the second, and I was able to look out of my eyes. The world was very blurry because both of my corneas had been cut with the laser. Trippy but no pain at all.

I laid down onto the second machine and the doctor worked on the right eye, then the left. For each eye, she cleaned it, then put the ring back in to keep the eye open, lifted the corneal flap made by the first machine, then lowered the second machine and told me to look at the green laser patch in the center of my visual field. The lasers didn't look like dots, they looked like the kinds of lasers you'd see in a light show - a green light show in the center of my vision and red light shows all around the edges.

Once things were aligned, the main lasers were activated and it took about 30 seconds per eye for them to resuface the cornea. Apparently each eye receives thousands of tiny laser adjustments, which is necessary to get rid of the various imperfections on the surface of the cornea. You can smell a little burning during this phase; it's a bit like the smell when they drill a tooth in the dentist office. But as usual, no pain or discomfort. One the main laser finished its job on a particular eye, the doctor lowered the flap and cleaned the eye.

It took about 20 minutes for the second machine to do its job, and then I had some transparent "eye shields" taped over my eyes to prevent accidental scratching or rubbing. It's important not to dislodge the corneal flaps during the first few days, since they are busy healing.

The entire process took about 2 hours, and I drove home in a cab. When I already look out of my eyes the world looked fairly clear but with a fine mist around everything - a bit like your vision if you go swimming with your eyes open in a swimming pool with chlorine. I went to bed and kept my eyes closed as they recommended. I've been sleeping most of the day and will go to sleep again after I eat dinner.

I can see my vision already getting better. I had a coffee on my terrace and watched the sun go down, and the sunset was fairly clear. Apparently the mist takes just a couple of days to go away, but it can take 1-3 months to experience the final result.

So far, everything seems to be going well. I have a follow-up appointment tomorrow, then another in a week, then a month, just to make sure it continues to go well.

Dog Drinking in Slo-Mo

I thought this was interesting; the dog's tongue goes the other way than was I was expecting.

Conscious vs. Subconscious

I often read articles that talk about the "conscious mind" vs. the "subconscious mind", as if they were two different areas of the brain and/or two different modes of thought. However, now that I've posted a metaphor for consciousness (the Conga dance metaphor), I can describe how I view the two terms.

As I mentioned previously, I believe that we become conscious of a thought as it gains in momentum. In the dance metaphor, the dance club becomes more conscious of the Conga as its gains more influence and as more people join it. So there isn't a "conscious mind", there are just growing structures that we become more conscious of. Similarly, there isn't a "subconscious mind", there are just structures that haven't grown powerful enough for us to become conscious of them.

To summarize, I think that ideas/thoughts/structures are reacting all the time in our mind, and some grow to the point that we become conscious of them. All the other reactions that haven't reached that point yet would be termed "subconscious".

Metaphors for Consciousness

Consciousness is one of those subjects that is difficult to talk about. So I'm always trying to find metaphors that can help me to explain my perspective on this complex subject.

One metaphor I came up with recently involves a dance club filled with people and a DJ. When the club first opens up, the people are milling around and there is very little dancing. As time passes, people start to dance and move from partner to partner as they try to find a little spark. Then, out of the blue, a couple of people start doing the Conga (a dance where people form into a long line). People around them cheer them on and some of them join the Conga. The line gradually grows longer, from 2 to 10 to 50 people. As the line grows longer, even the DJ notices and starts to play some suitable Latin music. When a hundred people have joined the Conga, the DJ says "Hey everyone, check out the Conga line!" Finally, the music stops, the people disperse, and the Conga line disappears.

In this metaphor, the dance club is a "mind" and the people are the "active structures". The people are forming and breaking bonds, and the Conga is a structure that grows from the bonding of many people. The "mind" gradually becomes conscious of the "Conga" and starts to influence more structures around it (i.e. other people start to join the Conga line). The DJ is the "language subsystem" that translates the active structure (the Conga) to a verbal representation ("Hey everyone, check out the Conga line!"). Note that the comment by the DJ is a consequence of the mind becoming conscious of the Conga, not the cause of consciousness. After the music stops, the bond between the structures breaks and the Conga line goes away. The "mind" is no longer conscious of the Conga.

I'll continue to blog metaphors for stuff like consciousness, because they are very useful for thinking about and discussing such a complex topic.

EDU 2.0 for Business

The "edu 2.0 for school" site has matured enough now that it's time to finally start working on the "edu 2.0 for business" version. This version will charge a monthly subscription fee and target businesses that want to provide online training to their employees/customers.

We're going to maintain a single code base that works for both the "school" and "business" versions, with conditional logic when required. My hope is that the conditional logic will be less than 0.01% of the code base.

The first step is simple culling of unneeded tabs and substitution of phrases. For example, "School" becomes "Business" and "teacher" becomes "instructor". In addition, we're about to release four new styles that are more "corporate" so it's easy for the users of the "school" and "business" versions of our system to choose something appropriate to their audience.

Here's a translated version of the home page that uses the default "school" style but adopts the new "business" nomenclature:

Consciousness

I visualize a mind as being a pool of active structures that are continually reacting with each other. 

For example, when you look at a face, your eyes pour a stream of tiny active structures that represent the smallest details of the visual scene into your mind. These structures partake in reactions that gradually clump various features together into higher and higher forms of structure, remaining active all the time. As parts of the face are recognized, the structure grows and includes portions that represent the mouth, the eyes, and nose, and so on. Finally, additional layers of structure that denote a face become involved in the 'reaction', which in turn start to react with structures related to the person whose face it is.

I visualize this process as if there was originally thousands of tiny distinct dots in a pool that gradually start to snap together and form a little pattern, which then combine with other little patterns to form an increasingly  complex structure floating in the pool. It's like watching a jigsaw puzzle assemble itself!

Which brings me to the notion of consciousness. 

The most common definition of consciousness is related to awareness. People will tend to say that they're conscious of something when they're aware of it. This can be applied to perception, such as simply viewing a beautiful sunset, or a higher-level thought such as "that was a great movie". Philosophers sometimes refer to these two categories of consciousness as "Phenomenal consciousness" and "Access consciousness", respectively.

My own understanding of consciousness is influenced by how I visualize the lowest-level workings of a mind. Specifically, I think that a mind's consciousness of something is proportionate to the richness of the active structures that are currently assembled for that thing.

For example, consider the example I gave related to seeing a face. When the input is first received, there are thousands of tiny active structures that are floating separately. Then after a short while, they start to clump together into larger structures which in turn form larger clumps. As the structure grows, we become more conscious of what it represents. 

There's a threshold at which a structure becomes large enough that it is translated into other forms for communication purposes (I discussed this in my last few posts). So the structure that represents an attractive face might be translated into the English sentence "that's a beautiful face" that we would then hear in our head. Note however that hearing the thought is a consequence of us being conscious of the face, not the cause.

The translated form feeds back into our sensory input machinery which in turn can cause additional growth in the original structure. In other words, generating translated forms of a structure can create a positive feedback loop that cause more of a mind's resources to be associated with the structure that triggered the translated form in the first place.

When we communicate a translated form of a structure to the outside world by, say, speaking about it, those around us might then start thinking about the same thing and a similar structure starts to form in their own minds. They in turn can speak the results of their thinking which can feed back into our own mind. Communication of translated thoughts to the outside world can therefore have the effect of bringing ever-expanding numbers of minds to bear upon a particular concept.

To summarize: I think that our consciousness of a particular thing is proportionate to the size and complexity of the active structure(s) that currently represent it. Consciousness is thus a sliding scale. Structures can grow and shrink in milliseconds, so what we are conscious of can change very quickly. The things we are most conscious of are often translated into phrases that we hear in our heads, but I want to emphasize again that we hear the thoughts because we have become conscious of them, not the other way around.

Coming up next: emotions.

Thoughts and Senses

In my previous post I likened a 'true' thought to an active 3D protein; in the same way that proteins react with each other to form more complex proteins, thoughts react with each other to produce more complex thoughts.

What about senses? When our visual cortex receives signals from our eyes, how are these represented and processed?

I believe that senses are processed in the same way as our thoughts. In fact, I consider them to be the same thing. I visualize the eyes as pouring in a stream of tiny active structures into our mind that immediately start reacting to form more complex active structures that represent higher level features of the visual scene. Ditto for the other sensory organs. 

As an aside, I think that the 'reaction' process for active structures is additive. For example, if two structures representing lines as different angles are recognized and bound by a third structure that represents a right angle, then all three structures are now clumped together as part of the reaction; the structure representing the right angle does not destroy the structures that represented the two lines. The concept of additive reactions will become very important later on in this series.

To summarize; I believe that pretty much all 'information' in a brain is represented as active structures that are continually reacting with each other. Some reactions give rise to increasingly complex assemblies of active structures, whereas others fizzle out into nothingness. Sensory input is immediately converted into thousands of active structures that represent the smallest features in the sensory domain, and these immediately start reacting to form larger active structures that represent higher-level features in the domain. 

So what I described in my last post as a 'true' thought is simply a non-trivial emergent active structure.

Thinking and Language, Part 2

I think it's useful to have a mental picture of how a mind works, and part of my mental picture relates to what I consider 'true' thoughts versus the 'translated' thoughts we hear in our head.

I think the relationship between a 'true' thought and a 'translated' thought is similar to that between a protein and its DNA sequence.

A protein is an active 3D structure that 'reacts' with other proteins around itself. These reactions sometimes form more complex proteins, and other times cause proteins to break down into smaller segments. A protein is neither "code" nor "data"; it's an active structure that has both qualities.

A DNA sequence on the other hand is a passive 2D encoding of a protein that is suitable for storage within the nucleus and may be easily written down.

I visualize our mind as containing hundreds of thousands of 'true' thoughts that are active 3D structures that react with one another, sometimes creating more complex thoughts and sometimes breaking thoughts down into smaller pieces. Now and again, our language system translates an active 3D thought into a passive 2D representation that is suitable for communication. It's this 'translated' thought which we hear in our head. Similarly, when we hear a sentence we can translate its passive 2D representation into an active 3D representation that can then react with the other active 3D thoughts in our mind.

I don't want to carry the metaphor too far because it oversimplifies what I think is really going on, but hopefully I've got across the idea that there are 'true' thoughts are active structures that react with each other and 'translated' thoughts are passive structures that don't react with anything.

Interestingly enough, the initial translation from an active form of a thought to a passive form is probably very fast - less than a second. The initial passive form would represent the complete sentence structure.

Then either the speech center can translate the sentence into a sequence of phonemes, or the writing center can translate the sentence into a sequence of pen movements.

Note that there's a parallel pipeline of translation occurring: active thought => passive translation of thought as grammatical structure => rendering of grammatical structure as speech or written word.

The inverse occurs when listening to someone speaking or reading a book. Each step can continue in parallel so that a partial representation of a sentence will have a corresponding partial active representation; otherwise you'd have to wait until the end of a sentence before processing it.