ΛορδΧηαος

This is the first post since I switched from Blogger to Wordpress. I had wanted to write a little something detailing my customization efforts with Wordpress using PHP and MySQL. I managed to put together my own multi-lingual theme (in case you haven’t noticed, the sidebar is fixed in Latin while the interface of a post will automatically match the language of the post), which was fun, and might be instructive to someone who is also trying to learn how to customize Wordpress……but alas I am not going to write about it now. Perhaps some other time.

What I want to write about, actually, is the mind.

In one of my previous articles lamenting about the meaning of life, I assumed that a dog does not ponder upon the self and its relation to the universe. I had said that strictly speaking we can’t really know, but the assumption is probably a safe bet.

However, that said nothing about whether a dog is intelligent.

Of course, it all depends on the definition of intelligence. A shit-eating dog barking at its own reflection in the mirror certainly does not seem intelligent (then again, some humans have been known to do the same…they call it “scatological fetish” and “self affirmation ritual”), but a trained trick-doing Doberman rolling over at the command of its owner does exhibit some degree of intelligence (ironically, when humans do the same we tend not to call it intelligent……).

So where do we draw the line? As a species, is canis lupus familiaris intelligent or not? Dogs can’t recognize themselves in the mirror, but chimpanzees can (i.e., they are self-ware), so chimps somehow possess a much higher level of intelligence compared to dogs. Yet chimps don’t solve differential equations, and they still eat their own shit.

It would seem that without a clear definition of intelligence, we can’t really go any further than what I was just doing. Oh sure, like looking for choir boys in a Catholic priest’s closet, you can dig up countless materials since Plato on definition of intelligence, with no shortage of detail, logic, and scholarly finesse…..all of which I am sure appear much more sophisticated than my talking about shit-eating animals. However, they are really just beating around the bush, and not hitting the proverbial spot.

The problem is actually this. Most of them define intelligence in terms of external behaviour. This line of thinking is best exemplified by the famous Turing Test. Obviously, the Turing Test is limited to humans conversing with machines, but it shows the paradigm of the behavioural view of intelligence: if a machine can fool you into thinking that you are talking to a human, then the machine’s behaviour (at least as far as conversation goes) is indistinguishable from a human; therefore who is to say that it does not possess human intelligence? Similarly, if intelligence is defined in terms of a set of behavioural traits, anything that exhibits a subset of these traits should be considered intelligent to some degree. The scale by which degrees of intelligence can be judged would be the percentage of the totality of intelligent traits an entity possesses. Homo sapiens would obviously be at the top of this scale, since we do not have any other way to judge intelligent traits except with the behaviours of human beings (scatological fetishes not withstanding).

Immediately, two obvious problems leap out like nude girls in Vatican. One, you can sit around staring into space and do absolutely nothing, exhibiting absolutely no intelligent behaviour, and yet could be thinking about the Schwarzschild’s solution to Einstein’s equation of general relativity, or nude girls in Vatican (either could be described as thinking and therefore can be loosely defined as intelligent). Sure, you might say, thinking is another behavioural trait, no?

Not so fast. Let’s imagine that a hairy blue intelligent blob from Alpha Centauri flies over and abducts a professor from Cambridge disabled with Amyotrophic Lateral Sclerosis but with an IQ of 800, without his computerized wheelchair, clothing, accessories, etc (since the transporter only beams up living organic materials, sort of like the time machine from the Terminator movies). The alien blob may easily consider its newly acquired naked specimen to be a strange sparse-haired monkey as no intelligent behaviour can be observed. Goodness help our professor when he screams in his mind “Don’t eat me, I’m intelligent and sentient! Look, I can count in prime numbers: 1 2 3 5 7 11….” while he is being fed into the food processor thingy.

Ah, wait a minute, you say. What if the Centaurian scans our planet first, and discovers that our professor is really not the norm of the species (which is sadly true anyway, or we’d be teaching quantum gravity to grade school children by now). In fact, there are buildings and roads all over the planet, and around the planet a certain range of the electromagnetic spectrum is bustling with communication. Our species is clearly intelligent, no?

Again, not so fast. This actually brings us to the second problem. This problem is best summarized by this following quote from Douglas Adams:

On the planet Earth, man had always assumed that he was more intelligent than dolphins because he had achieved so much… the wheel, New York, wars, and so on, whilst all the dolphins had ever done was muck about in the water having a good time. But conversely the dolphins believed themselves to be more intelligent than man for precisely the same reasons.

- from Hitchhiker’s Guide To The Galaxy

In short, having strange block structures, web-like connective paths, and other engineering byproducts littering all over the planet, does not constitute as an universal indication of intelligence. Ants and bees have the genetically pre-wired ability to build extensive hives, and yet we would be hard pressed to consider the species as being intelligent. Even if the alien is tuning directly into the correct range of the EM-spectrum where our global communication network occurs, an alien “intelligence” may not even recognize these “fuck you; fuck you too; fuck you three” electronic messages as intelligent in any conceivable way. They may use pheromones for face-to-face communication, while employing space-time ripples for long-range communication. Piggy-backing messages on top of atmospheric vibrations or electromagnetic waves may simply be something their brains cannot comprehend.

A third less obvious problem with defining intelligence in terms of behaviour is that it is not objectively quantifiable. Remember, if something can’t be quantified, it’s nothing more then philosophical bullshit (vis-a-vis the present article). How does one behaviour compare quantitatively against another? Is the ability to recognize unknown patterns more important (i.e., scores higher) than the ability to perform arithmatics? Subjectively we know we are more intelligent than chimps, because we can do much more than the chimps, but how much more? We also know we are more intelligent than computers because it is difficult for computers to adapt to new situations, even though they can out-perform us in repetitive calculations. We can do anything that computers can do, albeit at a much slower pace; on the other hand the computer clearly cannot do, at any speed, something we can do instantly, like recognizing a face at any distance, lighting condition, or viewing angle. Again, qualitatively we can say we are clearly more intelligent, but by how much? What exactly is the difference?

It would seem that a simplest way out of this conceptual dead-end is to ditch the behavioural paradigm of intelligence entirely.

I recently came across a conceptual model put forth by Jeff Hawkins (of the Palm/Treo fame), via his book “On Intelligence”. This model is quite simple and elegant. There is only one algorithm in the brain, he proposed, and the same algorithm is repeated at all levels in the cerebral cortex in all mammals. Intelligence is defined in terms of the presence and extend in which this algorithm is employed in the cerebral cortex. Even though Nature evolved the cerebral cortex to implement this algorithm, you do not need it to use this algorithm. However, the present incarnation of computers do not employ this algorithm, and therefore cannot be considered intelligent in any way at all despite their speed or software sophistication, even with massively parallel processing power like in the Earth Simulator.

The model is called the “Memory Prediction Framework”. The ubiquitous “algorithm” I just mentioned is a time-delayed auto-associative memory unit which remembers the sequence of a simple pattern in the signals, and has the ability to predict the next occurrence of the full pattern when presented with a partial signal. Each memory unit is composed of multiple columns of neurons in the cerebral cortex (each column is usually composed a few hundred neurons), and is arranged in a hierarchical structure. The lowest level handles raw input signals (by recognizing and predicting raw patterns), while higher levels recognize and predict patterns of patterns, etc.

The result of employing this algorithm at all levels of the brain’s hierarchical structure is that an invariant representation (a model based on patterns of patterns in the signal input) of the world is constructed in the higher level of the brain. We don’t need to remember a face in every angle, distance, or lighting conditions; through the feedback loop of the pattern prediction mechanism, the invariant model propagates from higher levels to lower levels and transformed back into specific patterns pertinent to the situation.

From this, all manners of the intelligent behaviours, consciousness, etc, can be understood and explained. To understand the model, we can remove the “hardware” component and simply talk about how the general algorithm works (which may be implemented by anything…neurons, neural networks, simulation software written in C++, whatever). Obviously to check whether our brains really do work this way, we still need to find biological evidence from neuroscience research.

Let me stop now, and promise to get into the model in a little bit more detail and discuss how, thanks to the crystal ball of our mind which is constantly making future predictions at all levels of cerebral functions, this conceptual framework can help us understand our own thought processes, emotions, and other aspects of our daily lives.