Artificial Intelligence, or AI is one of the most enduring themes of science fiction. How close are we? Depends what you want from AI. If you were to map the progress of Artificial Intelligence along the progression of a humans life, in many respects, AI is still a toddler playing with coloured blocks…
AI in the real world
If you want your car breaks to work when it’s wet, or your investments not to go south when a key market drops suddenly, then AI has those things covered. But if you’ve been reading Isaac Asimov and you fancy the idea of your own personal robot assistant, then you’re going to have to wait.
As enduring an idea AI is, we are probably many decades from something even close to a functional sentient cybernetic life form, capable of reasoning, forethought, insight and studious deeds & actions.
However, before those days come to pass, much work remains. And of those working hard to resolve the basic concepts behind more functional AI are the very smart people at MIT:
“Humans have a natural tendency to find order in sets of information, a skill that has proven difficult to replicate in computers. Faced with a large set of data, computers don’t know where to begin — unless they’re programmed to look for a specific structure, such as a hierarchy, linear order, or a set of clusters.”
As I’ve said before, the human brain is possibly the most advanced pattern recognition engine there is. When you think about it, our minds are constantly either looking for new patterns, or monitoring the progress of the cycles we live within, or measure our lives by.
Take time, for example, which is a succession of patterns within patterns. We measure the progress of time as series of units, such as seconds, hours, days, weeks, months and years, with each larger unit being comprised of a finite and fixed number of the preceding units.
To us, this is a simple system to grasp and is probably no less complex for a suitably equipped computer. However, what if we had a room with a random selection of twenty people in it — how do we organize those people?
That’s when things become measurably far more complex.
“Now, in an advance that may impact the field of artificial intelligence, a new model developed at MIT can help computers recognize patterns the same way that humans do. The model, reported earlier this month in the Proceedings of the National Academy of Science, can analyze a set of data and figure out which type of organizational structure best fits it.”
This work is huge step forward. Our understanding of patterns is borne out of simple observation; we’re readily capable of sorting through almost anything based entirely on the data supplied to us by our senses.
For a computer, this initial sorting process is where things invariably go horribly and hideously wrong.
“The model considers a range of possible data structures, such as trees, linear orders, rings, dominance hierarchies, clusters, etc. It finds the best-fitting structure of each type for a given data set and then picks the type of structure that best represents the data … Humans perform the same feat in everyday life, often unconsciously.”
Find a child of or around the age of five, sit them down, give them a box of coloured shapes and some time after they’ve got bored with playing with the box the shapes came in, they’ll begin to do something truly unique — the child will tend towards sorting the shapes by one pattern or another.
It can be something totally esoteric as colour or shape preference, but it’s something no computer has even slightest chance of replicating. From this starting point, you can see clearly the size of the challenge faced by these scientists, some of which are luminaries in their respective fields.
Right now, the most advanced AI software is barely as smart as a toddler when it comes to distinguishing patterns.
Visual challenges to AI
An icon of corporate identity design is the legendary IBM logo. This typeform logo is the three letters of the acronym, made up of horizontal lines. To us, this is simple to read. To a computer, they see only lines.
Similarly, any letters containing irregular lines, ligatures, or letters composed of other forms, such as dots or dashes are equally illegible. To anyone with the reading ability of a ten year-old, those words are most likely entirely readable.
In fact, such is the problem, these obfuscated letters & numbers are the very foundations of the CAPTCHA, a notoriously strong encryption system that befuddles rogue software intent on smattering comment spam on blogs like this one. A working example of a CAPTCHA can be found on my Contact page.
However, in an attempt to draw out every last remaining salable word on the web, Google have patented a system to read text within images. As an advance in technology, it’s a significant one. But if they have a working technology, then it’s a system that can be replicated by those intent on filling our blogs with utterly spurious and sometimes grossly offensive comments for profit.
So raise a glass for AI .. then drown the sorrows of bloggers the world over.
Artificial Intelligence — still looking for clues
To my mind, there’s nothing artificial about intelligence, be it the byproduct of a sentient life form, such as a human or a cetacean like a dolphin, or even a computer, if there is a true intelligence, then by definition it’s real.
In any case, semantics aside, we still have a room containing twenty people and computer only just coming to terms with a series of coloured, shaped blocks.
Our five year-old child could probably make a lot of sense from this room full of people. First of all, they could break our group down into some pretty obvious groupings, such a gender, race, age at a push and size with a few giggles and embarrassed expressions thrown in.
Our computer? Still playing the with coloured blocks, I’m afraid! This one’s a challenge too far. Worse still, our precocious five year-old child is now showing off; the man in the fireman’s uniform is obviously a fireman, the woman with the white coat and the stethoscope is obviously a doctor, and so on and so forth.
A voyage of self discovery
As you can see, the challenges faced by Artificial Intelligence are still colossal. But the future holds vast promise. Before we are able to equip any computer with the requisite wherewithal of a human, we must first know more about ourselves — and that to me is the great voyage that AI offers the human race.
We’ll be forced to delve deeper into our own psyche, peering into the very recesses of our consciousness, searching for the enigmatic source of the truly defining quality that separates humans from our fellow animals…