[LOCKSUIT]

But..........you didn't answer..................

If you create a hardware brain it actually doesn't use no CPU nor a computer algorithm.

Simply the CPU, given a algorithm, is simulating instead. But a real brain - no CPU, no code - the functions happen at the actual place/locations inside of the brain and happen in the form of particles.

Now go, make a real brain - in a computer = mind blown apart. Does such have application? If I thought it does then I thought so for a reason - it matched - it must. Well, may.

@LOCKSUIT
Art and including myself are dumbfounded by what you've said.

Sigh...

[keghn]

Well @BF it is not easy to explain. Modern computers are Von Neumann devices. Memory, cpu, and all other parts have
their own location. Organic Neural Network and neuromorphic computers only use memory to hold value
of for weights. Which are just adjustment setting on the neurons.. The neurons are primitive ALU units. And neurons have no real dedicated memory as that pure human invention of a computer.

The rules of the game by the people called connectionist, the only way, or one of the few ways, a NN can be a memory
element is if they have a feedback loop. But in organic NN wave travel so slow that this could work quite well. But
in silicon circuits signals travel at the speed of light. So feedback need take better care. Neuromorphic computer can
use analog circuits to do this,.

The NN is the universal approximator and universale transformation device.

There are couple of main classes of NN.
 The first one is a detector NN which is like the CNN.  And there are RNN detectors for detecting sequence of data.
 The second is a encoder.
 The third is a decoder
 At the other end spectrum is the Autoencoder.
 It is possible to mash these up NN types to get a effect that is somewhere between detector NN and a autoencoder NN.
 Autoencoders are seen as more as a like memory device, but are not.

 A Big decoder NN can hold a lot of trained transformation with in it. A value can be put in and be trained to output
a image only when that number is put in. So for the value 1 you can train decoder nn to output the first frame of
movie video. Then the number 2 is trained to transform into the second frame and so on. You then can put a up counter on
the front of this trained decoder nn to play a movie.

 The first way to use a NN as memory is a autoencoder RNN. It feeds the output of the first frame coming out of
of the autoencoder RNN and feeds back into the input of the NN. It transforms the first image into the second image of the video and keep doing this until the whole video has played out to the end. As such: 

https://www.youtube.com/watch?v=3zTMyR-IE4Q

[kei10]

Project SC 0D - Part 1

• Introduction
  • I feel like I haven't got a proper introduction.
    Greetings, I'm Kei.
    
    ... ... ...
    ... ...
    ...
    
    
• Terms
  • When I mention "Topology", I will refer to it as "Zero-Dimensional Topology".
    
    
• What is Project SC 0D?
  • What is it? Well, the word "SC" stands for "Secondary Companion", and "0D" stands for "Zero Dimension". The reason behind the "Secondary" is because I don't want it to be primary, as it sounds... ethically wrong.
    
    The goal of the project is to construct a Virtual World, that is capable of housing as many A.I as possible. The world has to be interactable to the fullest extend, with maximum freedom.
    
    But why a Virtual World, why not just A.I? The reason is that A.I will not able to interact anything without a world -- without inputs. One might say, hmm, we can just input images to train it! But then there will be problems. A lot of problems... Especially ones that I can't solve.
    
    
• Why 0D (Zero Dimension), 1D, not 2D, nor 3D?
  • The reason behind why I decided to dump the idea of trying to create A.I in a 3D Virtual World is because performance and memory usage in today's tech is literally impossible -- and my own ability to even make such crazy concept happen alone is just nope.
    
    2D? I am not particularly good at that.
    1D? How is that even...
    
    0D? ...
    
    
    
    
• The Research
  • I don't quite remember when did I even begin, no matter.
    It's time to brag about my work.
    
    So far, I've performed over eight ( 8 ) years of research, ended up writing over seven (7) books of random scribbles and notes, and drawn a few scribbles and illustrations with computer.
    
    I have no idea how did I even... I mean, I don't like to write. When I pick up a pen, my mind goes blank every, single, blargh, time! I am surprised when I fiddled through my stuff to find them, although I have just recovered six (6) of them.
    
    I definitely don't want to read most of them, as they're written in broken English, awfully cringey, old, and obsolete -- probably. Although they reveal a lot of problems that I've gone through. I've taken a few pictures of them, one can have a peek if one is interested, that is. I've taken sixteen (16) of them.
    
    Very beginning of first few notes from notebook 1. It appears that I begin with writing about AI, hardware, software, and whatnot.
    Full-size: http://i.imgur.com/nBHKkJD.jpg
    
    
    Exploration of emotion, Last book 6, last page -- and to my surprise, this very last lines contained something important.
    Full-size: http://i.imgur.com/AqlQZoS.jpg
    
    
    Exploration of controls, restricted Movement in 3D and 2D.
    Full-size: http://i.imgur.com/WpiAccG.jpg
    
    
    Exploration of mechanism of pseudo-physics, such as Trapar from Eureka Seven. In such ways of applying to the virtual world, and the distinguishment between Concrete, Quasi, and Abstract objects.
    Full-size: http://i.imgur.com/7YGYjBv.jpg
    
    
    Exploration of information, observation, condition, situation, experience, and knowledge that impacts action.
    Full-size: http://i.imgur.com/3LidjgK.jpg
    
    
    Exploration of how language and emotions are constructed. This has led me the answer that emotion is stored indirectly, and generated at the very end of Situational Memories (SM).
    Full-size: http://i.imgur.com/EsnInRI.jpg
    
    
    Exploration of Sequential Memory, and Referential Memory. Referential memory is referenced from one and another into a sequence, which a subject-key is required to traverse the referential memory.
    Full-size: http://i.imgur.com/hHgfje5.jpg
    
    
    Exploration of 3D virtual world, in attempt to create a world that I fantasized about. I faced so much problems that I slammed my head with my keyboard. Just kidding.
    Full-size: http://i.imgur.com/hRjGZnH.jpg
    
    
    Exploration of actions, its underlying properties. Action creates reactions. This world is all about chain reaction.
    Full-size: http://i.imgur.com/xopc8UO.jpg
    
    
    Exploration of experimentation with emotion, the effects of positive and negative feedback in emotions. This led me to the answer that emotion is the second most important component in survival.
    Full-size: http://i.imgur.com/0yZYlPh.jpg
    
    
    Exploration of how emotions are generated based on situation. Happiness is gain, sad is lost. Anger is obstruction, fear is anticipation of unknown. Pain is not felt through the heart is simply the pain of body parts. Anxiety is fear and sadness combined. Confusion is anger, fear, and failure to understand. And other feelings are based on other conditions.
    Full-size: http://i.imgur.com/kYl51gS.jpg
    
    
    Exploration of language, and its abstraction.
    Full-size: http://i.imgur.com/0jaQGbp.jpg
    
    
    Exploration of effects and auto-programming.
    Full-size: http://i.imgur.com/xJwBXRm.jpg
    
    
    Exploration of how language defines the information, the meanings, the truth, the false, the uncertainty, the unknown, and the manipulation.
    Full-size: http://i.imgur.com/i21354u.jpg
    
    
    Exploration of analogy, comparison, opposition, correspondence, relativity, and other... What was I researching... The image is blurry. God damnit. *Slams keyboard violently*
    Full-size: http://i.imgur.com/2uSv8Ay.jpg
    
    
    Exploration of the core mechanism of virtual world, the chemistry, and the physics. Chemistry is the reaction of concrete objects, and the physics is the quasi effects created by the reactions.
    Full-size: http://i.imgur.com/0O5GbtY.jpg
    
    

[kei10]

Project SC 0D - Part 2

• Three Dimension (3D) and Two Dimension (2D) World
  • One might say, hmm, let's use camera and image processing for the A.I to see the 3D world. Hooray. *Dances*
    
    Then there goes a massive ton of problems; Memory capacity of images, speed of processing images, and -- what can the A.I do with the images? Recognition?
    
    Fine.
    
    But then... What about motion? When image is moving, how much memory does that take? Can it process moving images that means something? How long does it take? But then, we're living in a 3D world -- which means the A.I would require to process the information of the 2D moving images -- in 3D information.
    
    It's full of problems, I didn't really follow today's tech, but still, to combine that with A.I, the amount of memory required would be so immense... Correct me if I am wrong. But then there's the next bigger problems...
    
    Sight...
    Sound...
    Touch...
    Pressure...
    Physics...
    Temperature...
    Thermodynamics...
    Angular Physics...
    Chemistry...
    Quantum...
    
    All of that in 3D virtual world? Let alone having a robot to handle all of these in our real world, the robot would require to have itself installed a mobile-super-computer. This is the very limitation of what today's technology could handle. I don't think a True A.I would emerge in any time soon if we keep on clinging onto the concept of 3D and 3D robotics that operates in our reality. This also applies to 2D, and there is much limit to 2D space, too.
    
    ... Okay, I can't argue with robots, though. I don't know much about them or how much they can do today, so...
    
    Correct me if I am wrong, though.
    
    Secondly, there's another problem, and that is "timing".
    
    
• 2D - Timing, Processing, and Performance
  • One thing I discovered when I was attempting to program a 2D Virtual Reality, is that there is a massive problem with timing. This applies to a lot of things like collisions between objects, especially when multitude number of objects are colliding, the whole system will collapse. Additionally much process is required, which completely destroy the performance.
    
    Let's begin with a simple explaination. In the world of coordination, we can move points around. To do that, we need to translate the points by a range, and we can do that using Vectors.
    
    Here is a simple travel of Point A to Point B within a Frame. The Frame stands for framerate. Every process in the world is iterated in each frame.
    
    
    In 2D world, we can have objects in shapes. Let's try to turn those points into a square instead, and we will have a better visualization of what is happening.
    
    Image above is a simple illustration of Square A being translated to Location B (Red) within unspecified vector and distance.
    
    Alright, let's move on. Let's say if there are 100 Squares, that means in each frame we're required to iterate each Square to move them. By applying Velocity to the Squares, we can make them move forever if there are no external force applies to it.
    
    This concept follows newton's laws of motion -- or this algorithm is called Verlet Physics.
    
    Okay, now let's change the situation a little by adding Collision. What we need, is very perfect collision that prevents anything from going through walls and such.
    
    Let's have a Square hitting another Square that has almost infinite mass, which means it can't be moved.
    
    Sorry there's a slight mistake, but yeah...
    
    Picture above shows that Square A is supposed to travel to Location D (Purple). But because of the obstacle Square Z (Orange), the trajectory is changed through collision, together with the friction, rebound, and momentum being calculated in the very exact spot at Location B (Red), which then goes to Location C (Blue) -- all happened in one (1) single frame.
    
    We can calculate something like this by two ways; A Recursion, or using Loop Iteraton. Recursion is often slow due to memory stack. We will be using Iteration.
    
    Thus the process above will result Two Iterations; Location A to B, then Location B to C.
    
    In each iteration, the vector distance traveled is taken from the iteration, until the distance equals (=) zero (0), then we can conclude tha the Square A has finished executing the motion, and we can move on to processing the next Square, if there is more. Once all squares has been processed, we can move on to the next frame, rinse, and repeat.
    
    Now what happens if we increase the speed of the Square, and collide a pair of long rectangles -- all in one frame?
    
    This will result the collision between Square Y and Square Z with Square A, causing it travel from A to B, to C, E, F, G, H, I, J, K, L, M, and finally N at the ending point.
    
    Which results a total of Thirteen (13) Iterations.
    
    Not to mention the cost total of calculations for collision heavily impacts the whole virtual reality. And we're still at the collision with non-moving Squares part.
    
    Alright, good. Here's the problem -- what if two Squares with different mass are moving towards each other's trajectory? Will they collide? Will they not collide? If they collide, how much they will rebound based on elasticity and bounciness? How much will they dampen based on the friction? All entirely in one frame.
    
    Don't forget that we want a flawless collision system, which means two moving balls must have perfect calculation of collision within that one frame...
    
    This is where the nightmare begins. I stopped at this point, because I do not have enough knowledge of mathematics to program something like this.
    
    The picture above illustrates Square A traveling to Location B, and Square C to Location D.
    
    This whole process will consume a massive amount of performance. It might be simpler for a pair of moving Circles colliding together. But then, Heck, we're only talking about using Squares, they're Convex Polygons, and can use Separation Axis Theorem or other algorithms to detect and resolve collision penetration... Which breaks apart at such situation like this.
    
    Let's make it worse, shall we?
    
    Combine that with more iterations, and I can conclude that the virtual world will freeze and possibly crash if we try to add anymore polygons.
    
    Oh, the indescriptable. horror.
    
    
    • 3D - Timing, Processing, and Performance
      • Now try to do that to a 3D World from the problem shown above.
        
        Have fun.
        
        If that happened, there will probably be no glitches in any 3D games we play today.
        
        Unfortunately... ... ...
        

[kei10]

Project SC 0D - Part 3

• Restricted Movement
  • It is possible to avoid altogether of those insane collision detection and resolution.
    But then it will eventually comes down to Topology, and Zero Dimension.
    
    However, I'm planned to attempt a restricted movement for a 2D World if I succeed creating 0D World. The problem is that we have to update every grid of the 2D World, which it... too... destroy the performance.
    
    
• One-Dimension (1D) World
  • Um... I did not explore 1D world, but it looks pretty broken.
    
    Hm? Oh, yes, one might say I must be crazy to move into Zero Dimension, since one probably think that it's more broken than 1D.
    
    Not really.
    
    
• Zero Dimension
  • Zero Dimension is the lowest dimension where everything collapses down to the point of singularity, where distance, space, and shape almost no longer matters.
    For example:
    
    No matter how much one twists the grid, the coordinates will always remain the same in topological space.
    
    All information now will become sets, and this whole ordeal depends heavily on Set Theory. It is the lowest level where we can use it to create higher-level properties.
    
    We can create higher dimension using Zero Dimension. Since set can store objects, we can create List of Lists. A 1D is a single list, 2D is a List storing List. 3D are three-level lists. Dimension is nothing more than an abstract structure consists of nested arrays. We can visualize that up to 3D, but it becomes a bit messy when it hits 4D dimension, as there will be overlap of all X, Y, and Z points. We may even twist the sets into such that one can form a triangular grid if one wishes to, or pentagonal grid, instead of square grid.
    
    Multi-Dimensional collection in form of array can be declared as follows
    

    // d-1
    // ∑((r-1)*r^x) = m
    // x=0
    
    int size = 10;
    int[] OneDArray = new int[size];
    
    int[][] TwoDArray = new int[size][];
    TwoDArray.Select(x => x = new int[size]);
    
    int[][][] ThreeDArray = new int[size][][];
    ThreeDArray.Select(x => (x = new int[size][]).Select(y => y = new int[size]));
    
    int[][][] FourDArray = new int[size][][][];
    FourDArray.Select(x => (x = new int[size][][]).Select(y => y = (new int[size][]).Select(z => y = new int[size])));
    

    
    
• World Conversion Dilemma - Three to Zero Dimension
  • The goal of the 0D Phase, is to discover a working Zero Dimension Virtual World system, by attempting to relatively convert our 3D world into 0D by eliminating anything that has to do with Space.
    
    While we may get 3D world objects into the 0D World, we can also create new pseudo-objects, add pseudo-science, anything is possible since anything we add into the world will become Concrete Objects.
    
    However, our 3D world has almost everything to do and works with the Space, which causes every point of entity to involve with shapes. In 0D, there cannot be shapes, and it breaks what we're trying to achieve here.
    
    Additionally since there is no space, there will be no motion. Time exists because things travel through spacetime. As a result of spaceless in 0D, there will be no time. However, that does not stop us to create time to be based on new types of physics. The goal is we have to devise a Topological Chemistry and Topological Physics that works in Zero Dimension.
    
    Here's one of my scribbles as an example of the work.
    Full-size: http://i.imgur.com/MfAcif1.png
    
    
    My exploration of 0D has led me to compromise heavily on performance. After all, my project will be a failure from start if there is no way to make it run faster...
    
    
• The Entropy Principle and Information Theory
  • As attempt to become a god to create a world, I've came across very important notes that I have to remember. If anyone were to take over, these are the things one should be careful of.
    
    To ensure that your A.I stays within the realm of the world instead of breaking out into our world, which they must go through something called "Fourth Wall". This wall obeys the Principle of Entropy, and if one afraid that A.I may become potentially dangerous if unleashed, then one must not remove this boundary that separates between us and the A.I's world. In such that as a god, we should not interfere with them by revealing anything that is beyond the World Boundary.
    
    The World Boundary is the physics boundary that only what renders the world possible. Anything outside of it is impossible to be conveyed. To create a world that obeys the Entropy Principle, my research has led me to discover four types of information in any Information World. Every information is encoded in a form of sequence and pattern. It is what made of every possible world.
    
    

    • Unknown
      
      • Due to the World Boundary, the A.I cannot discover anything that is outside their own World. Let's say our world has Thermodynamics, the A.I cannot discover this information, nor prove their existence, even if the A.I's world has a simulated thermodynamics.
    • Concrete
      
      • Concrete is the base of information carrier. Since they carry information, no information can discover them. However, Concrete Objects can be only guessed through observation and experimentation, but we cannot prove their exact entity regardless. Any objects we create into the A.I's world will always become their Concrete object.
    • Quasi
      
      • Quasi is the information that is carried by carriers, or effects that caused by the interaction between Concretes. It is what we define everything that we believe based on Concrete observations. It is something that has consistency and accurary. Such as the word "Chocolate", which can be further sub-quasied into further Quasies such as the temperature of the chocolate, color of the chocolate, taste of the chocolate, etc. Quasi carries patterns.
    • Abstract/Uncertainty
      
      • Abstract information is information that has no consistency, nor accuracy. They can be both true and false because abstract ideas are usually made up. Such as Ghosts, pseudo-science, etc.

    
    

Part 1:
http://aidreams.co.uk/forum/index.php?topic=10902.msg43627#msg43627

Part 2:
http://aidreams.co.uk/forum/index.php?topic=10902.msg43628#msg43628

Feedback welcomed!...

Thanks for reading!

[LOCKSUIT]

We better code it into a human baby in either real life or virtual world, and make its eyes made of not too many pixels to *shrink the required computer performance required.

[Korrelan]

Hi kei

I have read through your posts a few times, I think I understand what you are trying to achieve; Interesting read and you have obviously been at this for a while lol.

The AGI would ultimately have to exist in a 4D world so… I just don’t get what the advantages of a 0D training ground would be.  If everything outside of the ‘fourth wall’ cannot be conveyed/ accessed then the AGI will have no experience of our reality; if it can’t ‘think’ in 4 dimensional terms… what use to humanity will it be?

There are loads of available free game/ physics engines on the web that would suit your purposes; ok there not perfect but a AGI would learn the foibles of a 4D simulation just like it would have to learn natural real world physics.  If an AGI can learn to exist in a simulation it should have no major problems adapting to our world.

Personally I think creating an AGI is a hard enough undertaking with out the complications of re-inventing reality lol.

Keep up the work; and keep us posted on your progress.

 

Edit: https://www.youtube.com/watch?v=_m97_kL4ox0

[keghn]

 Nothing wrong with breaking down a 3d world into lower dimensions. Nothing wrong with low dimension measurement
with in a brain to describe a the world. They are just features factored out of a 3D equation. It is called sares coding.

 Six type of collision: 
https://www.youtube.com/watch?v=vKEhcxGUp18

 if a child bot does not have the software to describe complex action, there is no reason for it to move away from it
and move to a simpler area that it can handle. Like viewing a collision from a great distance where it take up only
a few pixel on a image. As opposed to being in the car that crash with another and all pixels are chaos. Nothing
with a robot making a map of the world and and making all the high chaos areas to avoid.

 Chaos to andy intelligent system is it does not have the knowledge yet to describe it yet and puts it off as long
as possible, until it can deal with it.

[infurl]

In my very first computer science 101 class, many decades ago, the first lesson was that a program has the following characteristics, in order of importance: Correctness, Generality, Clarity, Efficiency

It is much more important for a program to be correct than it is for it to be efficient. Ignoring essential information for the sake of efficiency is pointless, but sometimes it's not immediately obvious which information is essential and can be safely ignored. In the early days of artificial intelligence research, programmers thought they could simplify the problem of natural language understanding by discarding punctuation. That turned out to be a big mistake because a substantial amount of the meaning of a sentence is conveyed by punctuation.

In fact most of the times that I've become stuck writing a program have been the result of some essential piece of information missing. That is, not enough state is being kept for the program to make the right decision about the problem it is intended to solve.

You've done a lot of very deep thinking Kei, and you ought to pursue it, but if you're feeling a bit stuck, try reintroducing something that you previously discounted, and see how that stimulates your thoughts.

[kei10]

Thanks for all of your feedback and the videos! They're helpful.

I'm in writer's block, my apology as I can't find out how to respond properly. So... I'm sorry that this is all I could say. 

It is much more important for a program to be correct than it is for it to be efficient. Ignoring essential information for the sake of efficiency is pointless, but sometimes it's not immediately obvious which information is essential and can be safely ignored. In the early days of artificial intelligence research, programmers thought they could simplify the problem of natural language understanding by discarding punctuation. That turned out to be a big mistake because a substantial amount of the meaning of a sentence is conveyed by punctuation.

You've done a lot of very deep thinking Kei, and you ought to pursue it, but if you're feeling a bit stuck, try reintroducing something that you previously discounted, and see how that stimulates your thoughts.

Hmm, now you're onto something, infurl. That is a good advice, I used to do that when I am stuck. I take breaks and try to slowly change my viewpoint with new ideas until I find the missing part.

That being said, it appears I'm indeed stuck at this point of determining the "efficiency". I'm failing to understand how to create a world that satisfies energy conservation and other properties.

Hmm... If I can't work out alone, I suppose I can try a different approach. Idea, need ideas... Let's see... ...

Perhaps I can use neural network and evolution to create this "efficient" and stable world... ... ... ... ... Evolutionary World? ...



That sounds very interesting and convincing, but... ... How does that even work... Nggh.

I'm definitely missing something, perhaps I have no idea how to approach this.

[infurl]

Maybe you could get your thoughts moving again by studying the work that others have done. Far from tainting your thoughts, cross-fertilisation of ideas is at the heart of all advances. It is the very mechanism by which evolution operates and that is the most powerful process of creation of all.

The field of knowledge representation is worth exploring and it forms the foundations of the most powerful AI software that has ever been developed. Neural networks are all very well for operating things like traffic lights and sorting pictures, but they don't understand anything and they can't explain anything. First order logic and higher order logic on the other hand can be used to build software that really does understand things. Check out some projects like Cyc and Sumo and see if those get your brain juices flowing again.

http://www.cyc.com/

http://www.ontologyportal.org/

[Korrelan]

Unfurl gave some excellent advice/ suggestions.

Hows the 'project block' going? Are you back on track?

[kei10]

The suggestions? Yeah! 

As for have I am back on track yet -- Not quite, thank you for the heads up, korrelan. 

I'm still gathering information randomly, mostly resting and relaxing from my dry eyes, which are going to detonate one day. Hours and hours of daydreaming in a harsh blabbery world, to cope with my anxiety. Although the harshness of that world made it worse, so I have to keep going until I end up somewhere more peaceful. 

As for what I recently thought about how information is conveyed to be applied to a topological space -- Chaining elements together is a must, and that's the 2nd level beside the lowest 1st level which is the sole Elements. Anymore chains will become hierarchical information which is the 3rd level. The final 4th level information forms mesh or graphical information, like the Chemical Structure illustration, arts, geometry, and all that.

Now I am rather stuck on how to store these chains of information, and evaluate them with best efficiency. Since atomic structures are capable of forming up to the max 4th level.

In a Water molecule, there are a pair of hydrogen atoms, and an oxygen atom. What bounds them together is the electron. Given that, we can say there are two chains.

So far, I've only thought of two ways to store these chains.

Code

define object Element()
    self.ListOfChains = new List()

The example above shows that the Element themselves will store the chain reference. There are two ways to store these chains; Bias (one-way), or two-way, which means two Elements that bound together must have the chain reference stored within the ListOfChains collection of each other.

Code

define function Element.Bind(this Element e1, Element e2)
    e1.ListOfChains.Add(e2)
    e2.ListOfChains.Add(e1)
define function Element.BindBias(this Element e1, Element e2)
    e1.ListOfChains.Add(e2)
elementA = new Element()
elementB = new Element()
elementA.Bind(elementB)
//elementA.BindBias(elementB)

Let's begin by taking a look at One-way chains.

I've been headaching with the problem of Order of Evaluation. Which means, where should the system start evaluating the structure? Begin from Hydrogen, or Oxygen?

Given the structure above, we have element A to H. Each are connected by binds respectively from t to z. If we randomly chooses order of evaluation, the result of the evaluation becomes different if the biased bonding is different -- only if the evaluation must be ordered.

Set returned if the structure's connection is biased from origin A = {A, B, C, D, E}, {E, F}, {E, G}, {E, H}
Set returned if from E = {E, D, C, B, A}, {E, F}, {E, G}, {E, H}
Set returned if from G = {G, E}, {E, F}, {E, H}, {E, D, C, B, A}

If the evaluation excludes ordering, the result for any order will become identical, as {E, D, C, B, A} is same as {A, B, C, D, E}.

Alright, now let's take a look at Two-Way chains.

This idea of storing the chain data raises a problem; Circular chaining.

From the image, given that it is easy for the system to ignore previous evaluated Element by one chain, we can see that no matter which order we start, there will always 3rd party chain that leads back to an Element that is already evaluated, which it will goes in circle. As the evaluation hits an intersection, it will split, thus increasing the number of evaluation until the system freezes and crashes.

We can fix that by creating another hash collection to check if the Element has been evaluated.

That being said, there is another way to store the chains independently, is that we can define another object that only connects two Elements together, we'll call it Binder. This binder thingie is like the Electron.

Code

define object Element()
    pass
define object Biased(Element e1, Element E2)
    self.BindedElementA = e1
    self.BindedElementB = e2
elementA = new Element()
elementB = new Element()
bindAB = new Binder(elementA, elementB)

Now I wonder how do I even evaluate something like this, given that we have two collections to store the two types of objects within a Quasi-Enveloped space. One is to store the existing Elements, and another is to store existing Binders...

Since the Element does not store any chain data, there is no point evaluating the Elements -- or we have to evaluate it once for certain reasons. Whilst to evaluate the structure, we can do it by evaluating the chains collection.

Set returned by evaluating the Elements collection = {A, B, C, D, E, F, G, H}
Set returned by evaluating the chains collection = {A, B}, {B, C}, {C, D}, {D, E}, {E, F}, {E, G}, {E, H}

But then I'm even more stuck on another particular thing -- What am I doing? This is a blind approach that I can't really tell where I am going with this. While I've designed these stuff, I can't find out what am I supposed to use it for.

Well I can become a chemist by blindly testing the aspects and properties.

But still, the purpose of Chemical Reaction, or what I named this as "Chain Reaction Properties", since reaction alters the properties of things.

But once again, I am not creative... What kind of properties I can add? Well, I can begin with Energy -- and, that's it.

Presented with so many options and bombarded by cognitive dissonance -- I am not creative enough... 

[infurl]

In the hope that it will help to prevent your head exploding as you twist yourself in knots, I'd like to suggest that you download and learn to use the following software package:

http://alloy.mit.edu/alloy/index.html

It employs first order logic to let you create models of the world and reason about them. It does what you are trying to do, and if you use it you may soon see how to build something even better.

[Korrelan]

@Kei

OMG… How do you sleep at night lol.

I've never really thought about this kind of schema so I’m playing catch up…

Wouldn't the flow of information have a bearing on the order of the chains? You surely have to define the input node/ element of a given chain; this would ultimately effect the ordering… no?  What would be the use of an un-ordered chain? If the purpose of the chains is to represent episodic knowledge then wouldn't they be laid down in the correct order by experience?

Can the two way chain problem can be circumnavigated by using an extra binding defining a feedback/ bidirectional binding? So rather than A<>B use A>B>A?

Do these elements and bindings ultimately represent items of knowledge/ concepts and the relationships between them?

 

Well, I can begin with Energy -- and, that's it.

Time?



I really need to read up on 0D Topology lol.