On common sense

I should really be writting an essay for school about reality and virtual reality, but this was such a striking that I just had to write it down. It comes from a simple sentence in one of the articles leading up to my assignment, roughly translated from danish it said:

“Information has a form of control over the reciever. Of course you have your common sense to sort the good and the bad. However you do not have control over the information that is sent towards you.”

Then I began to think about the problem of common sense that I once looked at after watching this Google Tech Talk a few months back, and how the speaker in it argued about the fact how we waste the immense amount of information that gets poured into the virtual world, and how we could use human computing to build up a database of common sense. This database could than later be used to make our information better, hence, tell us the good from the bad better than we can!

I say better simply because we walk around in our isolated part of the world all day. Isolated by our identity(Prison Break, T-Bag, s4e12, near the end), by our work, community, government, shops, to hell with it, society. So although our brain has the ability to concieve so much more, we are bored with our isolated information, why do you think the world wide web is so popular?

Now, a bot that crawls the immensly wide web, could generate a just as wide common sense, or sense of good or bad, bias of it’s society, education and beliefs, simply because it has none. For the philosophers, this reminds of the state of the uknown, a part of the system for government of the unknown proposed by John Rawls.

The question simply lies in understanding the so wide web, that is written in such a wide natural language, by so widely intelligent human beings as ourselves. Basically, how do we get a machine to build up it’s own common sense? Or rather store it, building up can simply be done via human computing 🙂 The rest apperantly seems to follow with it?

An alternative search engine

I hate Google. Not because I think they are stupid, but because I believe that they are brilliant, or at least their search engine is. As a follow-up to the previous post, what Google does is provide us with the most popular information, which is partially based on how often the source is mentioned in other places on the web(links), partially how well it’s is presented and favoured by other users(page-ranking), and partially how relevant it is to the search keywords and your personality(the actual search algorithm, and monitoring of your search history).

So what Google does, is blindly tell us what we like, compared to what everybody else likes, thus defining our place in life and society, and we love Google for it. However, it does so in a rather decent manner, making that the only decent thing about it. The information is presented fast and clean, in theory presented as something bias. Only on the side or at the top of the page do you actually see the sponsored links, seperated, and simply formatted, indifferntiating from each other, except the catch-phrase and the name of the product, limiting the bureaus to only natural language to manipulate the consumer. In comparison to Times Square, this is perhaps as decent as it will get in the global generation, when it comes to information representation.

Nough said, there is no competition to Google. Yahoo is dead, and Live Search never got through puberty, Google own’s the web, and the web own’s what you know, and what you know own’s what you think, believe and live by. 

One thing important to understand about a single provider of a service, is that using it, becomes a religion. And when a phenomenon like religion occurs, the human mind dies, and when that happens, innovation stops, and we get a “middle age”, where we believe whatever we have is good enough, and don’t try to build new technologies.

The IT industry is exactly the opposite of this, you might argue, but then Google is as much a parasite in the system, as it is a building stone. Just think about it. You are SATISFIED with Google, a better solution seems firstly unneeded, and secodnly impossible, so you entrust all you privates to this giants’ posetion, and you feel good about it, becuase that just makes your search results better!

Well, let’s not get too synical, and abstract here, the world is full of conspiracy theories already, besides, they make your mind go ka-boom(whereas religion, as the sense of stability, calms it down). The simple point that I want to get across here is that someone needs to stand up to Google, and provide them some sort of competition, before we all end up like we have, with the white earbuds in our ears.

So how do we beat Google? Well breaking down a giant and rebeuilding it from the ground up is seldomly the answer, the best is always to innovate (a world all you marketing people are very fund of). Take whatever exists, with all it’s problems, and solve some of them. Where Google is weak, is accuracy. To find accurate and legitimate information via Google can be like finding a needle in a haystack, all depending on what sort of information it is. So we end up taking the top-most results, which are innevitably the most popular ones, but the most popular answer is not always the most accurate one.

This popularity aspect on the other side, brings us much closer to the global village that Marshall McLuhan so vitiously proclaimed. We can connect trends across continents thanks to Google, which provides same information to all of us, and we personally rate what is better and what is worse, and filter the bull from the truth. 

Einstein once said:

“I believe, there are 2 things that are infinite in this world – the universe and human stupidity, and I’m not sure about the first one.”

We as humans have a problem, and that is we are easy to manipulate to believe something sensational, popular, or trendy. That’s the way we went into any war in history, the most clear examples being Vietnam and Iraq, no real problem persisted, but we all believed it was the right thing to do, until proven otherwise. Please notice that it was not the governments that went to war, but us, because none of us could to see outside the box of society, beyond the popular trends, untill it was too late 🙂

So, what does this have to do with Google? Well, it means that Google provoked the war in Iraq, Bush told us one thing, and every top ranking page on Google said the same. Only further down could you ever find contradicting information, which we all know today was more accurate. At the same time, now that we have realised that we were mistaken, it is only thanks to Google that we have realised this, and have put Obama in the White House. So this popularity dimension has it’s pluses, but more importantly, it has it’s minuses, and this is the problem to tackle.

In the course of the past century we have all finally grown accustumed to the fact that the human mind is prominant to make mistakes, but somehow we relate that only to mathematical computations and leave complex decision-making up to the dumb human and it’s feelings. If we no longer trust the human mind to make the complex calculations, than why do we still rely upon the stupid mind to wheigh the different facts and their accuracy before we make an important decision?

A search engine that didn’t simply tell us what was most popular, but read, understood, weighed the information, and presented it in a decent, relevant to us manner, so that we then as people could make most rational decisions based on accuracy, and not on simple popular belief. Imagine how many wars could be avoided, how much new successful business could be built, how many market crashes could be avoided, and what a better place the world would be to live in!

Such a search engine seems distant, but it isn’t quite as far away as we may think. It is only half an AI, since it hasn’t to make any real decisions, but simply sort everything that has already been decided, into right and wrong decisions, not simply popular and unpopular truth. Then it can levae it up to us, to believe this machine or not.

The question then of course is, why  would we believe it then? We don’t trust a machine to even drive our cars, why would we entrust it our political decisions? Well, imagine the indexing capacity of this monster, it is the slave, no human could ever be made into, or could duplicate. Reading the immense amounts of information fed into the internet 24 hours a day, 7 days a week. If done right, this guy can make rational explanations way beyond any human’s possible ability, simply because of the bigger “brain capacity”.

There is another aspect to this search engine that makes it brilliant, and that is the fact that it actually helps us make decisions, something we are becoming extremely bad at because of the huge amount of choices and decisions that have to be made, our mind is simply way to occupied with the everyday decisions to make right choices at work, or perhaps the other way around, depending on the person of the case. Explaining the difficulty of this sea of choices would take another 1000 words, so I simply suggest watching this Google talk instead, and imagine thereafter how such a decision-sorting engine could help, if it knew enough about the world and you: http://video.google.com/videoplay?docid=6127548813950043200&ei=5ZskSaGnE6f22gKj1LX8Bg&q=choice+google+tech+talk

No great conclusion for you guys this time, enough to consider already 🙂

Ctrl+Z

Taking verbs out of the scope of dimensions, and regarding them as isolated descriptions of the relationships means that adverbs, adjectives, and alike, also have to be isolated in the same manner. That starts looking extremely complex. Secondly, this makes no sense what-so-ever. The fact of the matter is, those relationships exist across many different dimensions. As a simple and streight forward example – many different things can brake, that aligns all those objects on a simple dimension of things that can brake. 

I guess that was an example of those misunderstandings my mind goes through every time it sees something nice and shiny, or learns something new, I guess I’m easily seducted, and that’s a bad characteristic to have in the IT world, or maybe not??

Anyways, everything lies in dimensions, and is a dimension by itself. That means that the semantic net must be a multi-dimensional data store, kind of like friends on Facebook. Question is how to implement it?

NLP aka. AI

Well, it’s been a while since I posted something, but I guess I’ve been thinking..

Thinking about my highschool project, now named, Natural Language Processing(which changed the goal, as a general, quite a bit). After a few weeks of research, I’ve become familiar with the known problems of bringing NLP to life, and I must admit, these problems look much alike problems within machine visual, aquistic, taste and sensory recognition. This really leaves me wondering, why are these areas of research, regarded as so completely different, when all of them basically require – observing, storing, prioritising and querying the data?(Excuse me if I missed some, but you get my point 🙂 ).

Why are we trying to achieve each one of these goals on their own(where they do seem extremely complicated), instead of creating a unified solution which can be used across the different areas of concerns? I guess the reason for this is the frightful name of such a solution -> artificial intelligence.

However, if you look at the world one detail at a time, as we do with the different aspects of AI, it seems extremely complex and chaotic, if we could find one simple truth, one simple rule, everything becomes simple and harmonic. In just the same way I have come to believe that recognising words and language is exactly the same as recognising shapes and colors, the only difference is the observing channel, what we have now is different terms for essentially the same thing, and that, in my mind, seems utterly stupid.

This leads me to my scary conclusion -> The easiest way to make something that trully can understand and speak like a human, is to create something that is fundamentally same as a human, a so-called artificial intelligence.

On babies

Another aspect that concerns me in AI development, is how advanced we expect the systems to be already at their starting point.

It is simple to understand, that a 1-year-old baby cannot understand a Stanford professor, first it needs to learn, observe and understand, and it must do so in a context close enough to the context of the Stanford professor, if we want it follow along the same lines of thought as the professor. So, why is it so tough to understand, that without having an idea of the world, the AI cannot come with complete and meaningful to us answers?

What we do nowadays is create agents that take a read through the Oxford dictionary, create a semantic net of the meanings of words in the English language, and rely on statistical calculations for everything else. This is like locking up a kid in basement and forcing him/her to read and memorise the dictionary. That creates a relatively good and smart source of information, but it has no understanding outside of that dictionary, which scares us, we are social beings, and send people that lock their kids up to prison(well most societies do at least). And one small detail, how does the kid know how to read in the first place?

Recently a Googler had posted on the official Google blog, that she belived that Google has acieved a rough 90% efficiency in finding what we want, however the last 10% require 90% of the work, and here she’s essentially talking about correctly and ligitimately understanding the human and the knowledge that we’ve preserved on the world wide web. However NLP in information querying is a whole other post 🙂

So perhaps babies hold the answer to the worlds problems. If developers want to create a true AI, they should create babies, not professors 🙂

Dimensions, Points and Natural Language

So what the hell do all these dimensions and points have to do with natural language? Well let’s try to analyze a few sentences.

The dog is green.

The dog is an object, and is therefore imediately referenced to a rough synth of a dog in the 3 dimensions of space. The word is indicates a reference to some sort of dimension of function or property, and the word green is a reference to a point in dimension of color. If throughout the dimension of time there hasn’t been a case of a dog being green, than no such connection will be drawn and the idea of a green dog should therefore seem odd to the intelligence. 

This, however, does not stop the intelligence from imagening such a dog, meaning that an intelligence is seemlessly able to mix dimensions as it wishes, if this is a realistic mix, we call this an act of creativity, since as defined, a green dog has not been seen before. If such a mix of dimensions can serve to solve some sort of a problem, we call this innovation, and is often where we dicide to earn some money on our idea.

So.. I’d like to define knowledge as knowing the broadth of diverse dimensions, learning as aquiring that knowledge through observation and imagination, imagination, in turn, as the act of creatively mixing dimensions, and finally thought as the act of analysis of what mixing a few dimensions would result in. Whether you pick the right choice through thought or not, depends on, basically, your previously acquired knowledge and, well, luck. Since it is knowledge you acquire from this imagination, the quality of that knowledge depends on the quality of your choices of the mixture, which depend on the previously aquired knowledge.

So in essence all the information should saved in these dimensions, and what is contained within these dimensions are relationships to all other possible intertwining dimensions at the specific points of the dimension.

So here I have analyzed how to store the analyzed data, i still need to know how to weigh acquisition and usage of it.

P.S. don’t be mistaken by the term knowledge, it has no direct relation to ONLY scientific knowledge, it also relates to knowledge about mixture of colors, seeds or even actions and sounds for that matter.

The Dimensionz

I described earlier that there must be some basic dimensions, such as shape, function and sound, but have come to a conclusion, that sounds way too stupid. 

First of, shape, is nothing else than elements of color twisted in the 3 dimensions of space, and color is nothing else than elements of light twisted in some other higher dimensions, and so on, and so forth. Same goes for function and sound, although those 3 sure look like as some very basic dimensions, we can go deeper, so they are merely MAIN dimensions, which will be used most often, but are not at all the building blocks.

They are, as I myself defined, dimensions that can be twisted and indexed as pretty much everything else. Instead the building blocks seem to be the points themselves, and the connections that they have. The building blocks of our understannding of the world seem to be built up of:

New Dimension(IndexOfSubDimension,IndeciesInSubDimension[])

I was too lazy to desribe this in NL, which is another plus side of regular PL’s which I should look into. 

A simple example is a pencil. In the dimension of time it appears in the past, now and most likely in the future, if you are holding one in your hand. 

However I have defined that everything is a dimension, so how is a pencil a dimension? Well it isn’t, not in the current context of thought. The central point of the context of thought would most likely be a point, which in our cas is a pencil, but can probably be twisted to be a dimension in another context. Think of this like the 2 basic dimensions, sth in the first dimension is a point, but once you add the other, you can make it a line. Or well, almost like that.

However a pencil is still a dimension, becuase it has a dynamic number of properties and functions, a different number of points so to say.

Natural Languagez

So in one weekend I managed to drink a whole lot of free beer with a bunch of freshman on the datalogistics course at Århus University, visit Baresso in Århus the morning after, go to my ex-girlfriend’s birthday party in Grenå(just to score 4/5 phonenumbers, and a few kisses), go to Copenhagen the morning after to hand in my application for a permanent residence in DK, and sit down at a Baresso on Kongens Nytorv and write this entry(I’ve missed out a few things, just so you don’t get borred). 

Despite the busy schedule(which I had, well, no idea about Friday morning, which shows you the spontaneity of my choices 🙂 ), I still managed to get a thorough read of the introductory and concluding chapter in “Human Language And Our Reptilian Brain The Subcortical Bases Of Speech Syntax and thought“, by Philipe Lieberman. This, together with some previous knowledge and research, as well as a talk with Bjarke Bondo(on our way to free beer, and a presentation from V8 developers, which we unfortunately missed), and my mom last night, has given me an idea of where to steer my further line of code vs. line of thought research. In fact I once again believe that I have a theory that looks so elegant, that it would bring AI to the world, make world piece and show us the deepest secrets of the human mind, but I’m probably mistaken, and this is exactly what this blog is for, I want to see how often, and how far off some of my thoughts are. So here it goes, all or nothing: 🙂

The thing with natural language is that to many eyes it looks way too complicated and way to differential and abstracr, that there can impossibly be a set of rules like there is for regular PL’s. Naom Chomsky argues though that since we all(by default) have the ability to fall into any languistic syntax and understand it, than there must be some sort of a Universal Grammar embedded within our genes. This idea has taken a head start about 50 years ago, and has bothered many neurologists, linguists and alike, for years, but noone has been able to come up with a unifying and overall syntax, for anything at all.

• One problem with this research is of course that it has absolutely no respect to possible future developments of natural language, and only looks at past experiences. 
• Another problem with the research done in this area is the question of how these words are stored within our minds? Do we relate shapes to actions to sounds(and round we go)?
• Than there are of course those that believe that there is no unification and overall formula for the syntax of NL, and that everything is pretty random, but as I described earlier, I don’t believe in random 🙂 the underlying idea of random is way too weak, and has been proven wrong too many times. So those who believe in random, are by definition insane. Since they look at the same action and expect a different result(same goes for the act of love).

If we try to look at our own minds, then our thoughts will most likely be in a form of NL, and one that is most actual in the context of the thought. However simoultaneously thoughts of pictures break down this idea. So sounds are intertvined with forms and all are intertvined with actions(& consequences, which are the same thing in this context). All of these seem to be blended together into some sort of synth, alike Microsoft Photosynth, only having MANY MANY MANY more than the 3 dimensions of space. So imagine the brain like the 3D dot-representation of the real world only it, is in roughly ∞ amount of artificial dimensions, the amount of which grow and shrink depending on what our thought wants(in essense the 3 dimensions of space are too AD’s, we are just so used to using them, that we are blinded by it)

So right… if we look aside from all the extra dimensions, then a single entity(or so-called object) seems like a point in the 3 dimensions of space. Let’s think of a pencil, hovering in mid-air, and then add on a dimension of time, here, the simple point of a pencil will most likely appear many times, rather than once, because we have used it in many different locations. Does this mean that for every occurance in time there is a seperate point in the 3 dimensions of space referencing the pencil? Although simple, but most likely not.. So instead I like to believe that the pencil is located in our simple dimension of shapes, than every single of the 3 dimensions of space are bended by the context of thought, at this point the context being a point in time. In essense the 3 dimensions also bend each other – try rotating the pencil in your head, the object’s shape in the terms of x and y depend heavily on it’s position in z. 

This makes the idea of the overall picture of something quite dynamic as well, since that simply depends on the amount of dimensions that you wish look at, and whether your understand the point’s position in all of the dimensions, if not, you get confused, like you were confused over the placement of the letter z in the AD of my blog, but you knew that it was a letter in the gregorian alphabet and knew what it sounded like.

Right, let’s abandon the thought of dimensions bending each other for now, and try to look at this at a larger scale. If we regard natural language as a dimension of itself, it will only make sense if we know it’s position in the overall picture of whatever it is that we want to describe. So for an AI to understand our thoughts, it must know the flow of our thoughts, and thereby know us, and the better it knows our thoughts, the better it will be able to understand and interprate whatever it is that we are telling it.

Understanding the individual is not enough though, and is thereafter nearly impossible. It will be much easier for the AI to understand the individual if it has a picture of the society, and the overall picture on the same scale as the whole of the human kind. Afterall, it is the society around us that defines, what, how, when and why we do whatever. 

So in order for an AI to become an AI it must be in the cloud, and it must be personalized for each user, kind of like personalised Google search. It learns your habits, and the habits of those in your community(as well as the whole picture), and makes the best possible guess on what it really is that you want. 

My thought therefore is, if we can get the machine to guess what information we need in the current context, than why not try and make it guess what function it is that we want? -> Programming with natural language

So what we get on the personal level, is an AI that knows us, and can help us with our tasks and choices, but on the broader level, connecting these AI’s creates an AI society. The properties that they can share, like a language, are common for that society. Combining these AI’s on a global level, and using computing to determine global values between all humans, could be called a Rational Intelligence(RI). Rational because this intelligence will most likely have a very good idea of all possible broadths of dimensions known to all men, and not have positions and opinions of it’s own. This in turn can serve as the soul basis for a global government. 

The only question is, how this AI weighs the different points and dimensions.. To be continued..