From the New Scientist article Daydream Believers (from last week):
“I would love to know if chimpanzees can entertain the notion of a unicorn, but we have no idea,” he says. “As far as I can tell, we don’t even know whether they can entertain two possible scenarios to solve a problem.” In Call’s view, it is impossible to say whether the animals that solve problems without trial and error are consciously imagining different solutions, or subconsciously integrating information to come up with the correct solution. “I’m not saying animals can’t imagine two different scenarios,” he says. “I just don’t see the evidence for it.”
As I have been learning this week from AI experts, the ability to learn depends mostly on trial and error processes which strengthen neural pathways that lead to positive results.
But as I am also learning, the field of artificial intelligence is split between three different types of AI learning processes.
First there are big data systems. These aren’t really intelligent because they depend on pre-programming by intelligent agents. They’re effective only because of brute force approaches to problems, processing huge amounts of data. You can imagine the problem. It’s a very energy intensive way to achieve the optimum action, and it remains dependent on instructions on what to look for. Like finding a needle in a haystack by processing every single bit of hay and checking that it isn’t by chance a needle. This is brute force trial and error.
Then there’s proper AI which achieves intelligence through self-learning algorithms that reward the AI for reaching objectives of their own accord. These are a different type of trial and error process. You don’t have to know what the task is, you’re simply learning to deal with the environment you are faced with and every time you discover something exploitable you develop a neural pathway that teaches you this is a good way to approach that sort of problem.
Problem with this system is that, if an unexpected obstacle hits the AI (unexpected item in the bagging area), it can get stuck in blind alleys applying approaches it has found effective in the past until it finds a work around. This can take a while. But sometimes it may not work at all.
The third system — arguably the smartest — incorporates what i have come to understand as mental visualization/modeling. It’s a system that aims to replicate the way the human brain works.
This is achieved by the AI integrating a model of its perceived surroundings into its actions, so if and when it comes across an unexpected obstacle it can revert to its model and take a good intuitive guess about where the solution might be. In other words it uses its imagination to figure out a solution.
The more creative it is the more likely it is to think outside the proverbial box and find the sort of solution that a process-driven AI would never be able to imagine.
A lot of human intelligence thus seems to be linked to the ability to create a working internal model of surroundings and to be able to cut corners through processes which would otherwise take ages and/or loads of energy.
But I do wonder if perhaps we have inherited a dual system as reflected by the presence of both process-type thinkers and creative-type thinkers in society.
The latter of course are pattern recognizers and extrapolators. These are people who have a tendency to jump to conclusions based on instinct and hunches, without too much care for the process of how to arrive at that solution. These are the sort of people who are super intuitive, who at school didn’t have to study too hard to pass exams, because they found they could get away with getting good marks with only half the amount of work of everyone else.
The former, however, are dedicated to the process. If you provide them with an unexpected solution to a problem they won’t necessarily trust you until they have worked out for themselves the process by which you can arrive at that point. These are, in other words, book learners. At school they would have been the type that got incredibly good results but had to study hard (and probably enjoyed that studying) so as to get to that end result.
There are, of course, strengths and weaknesses in both systems.
The amazing strength of book learners (I.e. non visualization people) is that once they understand the process they become better than anyone at doing it. They can become human calculators. And are generally hugely efficient at whatever skill they learn, and don’t get bored easily.
Their weakness, however, is that they can get stuck in the little picture and go down blind alleys. Also, they tend to be scared of unexpected encounters or anything they are not used to. In the extreme they probably fall on the autistic side of the psychological spectrum. On the really extreme side they may have trouble recognizing faces and emotions because of poor pattern recognition skills.
The amazing strength of creatives (people who can visualize from models) is that they can cut corners and find the most unexpected solution to problems and obstacles. That said they’re probably only efficient for short periods of time, with the efficiency highly linked to the out-of-box thinking process that leads them to an unexpected outcome. This may encourage an obsessive personality, I would imagine, which labours to prove that its pathway is correct. Nevertheless, they are very good at spotting patterns and are probably very good at visual representation, the arts and the subtleties of tones etc due to the power of their mind’s eye.
The downside, however, is that they do get bored easily by everything not related to their current preoccupation and can get carried away by their imaginations, never bothering to back up or update their models with actual confirmed processes from elsewhere. If they become too used to extrapolating, it can probably lead to extreme eccentricity and/or madness. Also, just because you can model internally doesn’t mean the model is necessarily correct. In the extreme scenario, you can imagine how eccentric some of these people may get and how much they may end up disregarding process, to the point of becoming non-functional or disruptive members of society. It seems quite likely to me (extrapolating madly, of course) that dyslexic types fall into this category.
The biggest advantage of the creative mind, however, is that it is labour-saving in terms of how many trial and error processes one has to go through to reach a conclusion. Consequently it is much more efficient, albeit not necessarily as accurate as the process-driven mind.
I imagine in society we need a good distribution of both types of brains because in the long run both are needed and both complement each other.
Give a creative mind a well formed model (I.e. One that has a wide range of data points/processes to draw on) and you get to some very special intelligence. And that presumably is the real potential of an artificial intelligence which has the capacity to model (visualize) but also has at its fingertips an expansive data set to feed those models. You go from brute force analysis of everything towards achieving action through dynamic real-time extrapolation that is more accurate than usual.
But I’ll have more on this, as well as the hierarchal architecture of the brain (which apparently features a fascinating competitive neural structure reminiscent of evolutionary game theory and competitive forces more generally) , when I actually sit down and read the papers which were given to me.
Quick update. I think all the above explains the success of the Myers Briggs system. Not sure how Introvert vs extrovert comes into it, but intuitive vs sensing does make a lot of sense. A sensing agent is the process-driven one, while the intuitive one has a good working model. Thinking vs feeling probably reflects our ability to back up that intuition with fact. As an ENFP, I just run with whacky extrapolations rather than using that intuition to figure out the process. I just “feel” it must be right. Thankfully I’m only marginally feeling, so I do have a lot of respect for confirmation and if the facts prove my intuition wrong I will change my mind (begrudgingly!). If you are intuitive and extremely feeling you can imagine a tendency towards cultism. Last there is perceiving vs judging. I guess that applies to how you respond to extrapolated options. Interestingly I have a tendency to orientate towards other ENFPs and/or INTJ/P types in life. I can see how this is complementary. INTs probably have extremely good intuitive models so can understand what crazies like ENFPs are on about, but their respect for process means they are not only great at deducing solutions they are also great at proving them. INTJs tend to be rude with an ability to alienate other personality types (other than ENFPs). Small surprise INTJs are the least common personality type. INTPs are also rare.
I think it’s pretty likely that if we do create an AI it will have the personality of an INTsomething.
ENFPs (my personality type) are all extreme extrapolators. Namely creative writers, journalists, idealists and err in some cases conspiracy theorists.
Although, obviously, hard to know how true those classifications are (especially for the historic figures).
But I do think it’s interesting that logic suggests the extreme opposite of the above should be an ISFJ type. Logic further suggests these types should be dedicated to process-focused actions and/or good at following orders. Funnily enough a) there aren’t as many famous ISFJs as other types, and those who are famous tend to be military people.
Process and brute force accompanied by “thinking” rather than “feeling” meanwhile allows for very effective, convincing and powerful leaders to emerge.