[keghn]

 To focus on one thing and let everything else everything else go in it random way.     

 The focus is a pinpoint tracking program. 

 Once the focus has figured out a pattern. It is given to the automated sub focusing tracker. 

 Hundreds of auto tracking routines can be stored or learned. But only one Main focus can be running at one time. 

 The main focus can start a existing automated focuses accounts and can up to hundreds running at the same time. 
 If a automated focus destabilizes then the main focus take over to shore it up. 

 For example you learn to walk as a child. The main focus at work. 
 And now you do not even think about it while the main focus is learning how to juggle. 
 If slip then main focus takes over walking.

 The main focus aso orchestrates and make auto orchestrater routines.   

[keghn]

 
 The main focus is consciousness. The automated focus is subconsciousness.

[LOCKSUIT]

You mean, the brain is always selecting actions/senses 1 at a time, but many many many can be self-igniting on their own. Ex. a build-up - walk, use phone, whistle, grind teeth, and now you are doing 4 things at once.

I understand what you mean.

[keghn]

  The AGI first plans a movement. Which as a baby, right now. Does not know how to do . But has seen some else do it.
Plans on moving its arm up and down. This will take a lot cpu power to figure out the movement. Once it is figures out, it just a recording that can be replayed by a lesser focus.
The main focus program can go onto new thing. Like move the hand up and down. Once that is figured out by
the main focus the sub focus can run it.
 The main focus can also start and stops and combined two or more  already learned movements.
like moving the arm up and down and the hand at the some time.

 Just like music, the Main focus uses a lot of cpu power to turns a five second noise into sound.
 Once it turns that noise into a sound, that it like, it is save for later us. To be combined with other sound samples.

 A machine use  cnc go code to play out machine movements of it robotic arm.
CAD programs can run auto listp to do auto drafting? 

[Zero]

Imagining, perceiving and acting rely on the same "summarizing" engine.

When you learn how to drive a car, you first have to be really focused on what to do, because nothing is automated yet.

But after repeating the same small actions again and again together, these low-level actions get summarized by a single higher-level action, or rather several higher-level configurable actions, that can be run autonomously. It's like functions definition, really.

Then when you're used to drive, you can drive and talk to someone at the same time, because the driving engine is almost entirely autonomous. Only willingness and unexpected events will draw attention back to the road, with heavy focus.

While multiple tasks are running, focus often switch from one to another, not to act at low-level, but to pilot the acting engine, which is like modifying the parameters you give to a function.

Wikipedia, as always, is a good resource on the topic.

EDIT: Does the way I describe things here corresponds to the way you see/understand focus?

[LOCKSUIT]

Generate actions or copy actions. Learning actions by rewards.

Then using them by matches. Self-ignitions can happen while you add more matches into the active playlist.

Walking while talking while clapping.

Where'd the actions come from? And how are all 3 being done at once? Answer is above.

[keghn]

 I am thinking out loud about machine consciousness. 
 There is only one main focus. If for some reason
there is more than one Main Focus they most likely end up fighting each other and AGI will go into a state of mental
disorder. There only can be one! There is only one queen bee. All others are drones.
 There are many many lessor focuses.
 When the Main focus start a old pattern by giving it to a lessor focus and tell it to run it.  The main focus has information and makes
note on this pattern. like how long to let it run, run time, set a timer to come back and check on it. or wait   
 for the lessor focus to scream a alert that that it is loosing control or it not going as predicted. 

 The information in these replayed patterns are output motor controls instructions, volume levels, and input data patterns   
that are expected to come in from the sensors of the, ears, eyes, and motor positions. 

 If the AGI is in a state of emergency the main focus will play old pattern that are safe and sure and do nothing creative. it will
keep a close watch on the lessor focuses for the slightest alert. 
 When The AGI is in a save place it will be trying out new pattern combinations. Or make a completely new patterns .

[ivan.moony]

The main focus could be a thread manager for sub-focuses.

[Korrelan]

Interesting topic…

Attention/ focus / saliency/ etc is what guides/ links/ chains the individual thought facets together to create an complete thought/ memory/ action.  ‘Focus’ as we call it has many variations; even touching all your fingers individually to your thumb requires focus/ attention at many different levels.

An analogy: Topic attention and focus are like a constantly moving landscape of hills and valleys that guide the flow of the river/ consciousness.  Imagine experience as pre-carved tributaries and attention/ prediction influencing the landscape ahead of the flow; pre-charged/ semi-activated networks poised to fire depending on the current input streams. At any point along the flows journey there are multiple valleys of choice… a dry riverbed is a prediction of where the water can/ might flow.  The actual path the water eventually takes is a property of the water not the landscape... top down/ bottom up.

Focus is an inherent part of the knowledge representation in the connectome.  Look at this picture; it’s just like any other JPG but it literally has focus (not the word lol)… the focus of the picture is inherent in the data representation.  This pattern of pixels creates a point of clarity/ meaning…you can mentally focus your attention on any part of the image but the only section that you acquire a ‘pattern lock’ is the center.



Information/ knowledge is stored in the cortex/ connectome with ‘focus’ already embedded in the data schema… because that’s what you where focusing on when you experienced/ stored it.

 

[Zero]

An analogy: Topic attention and focus are like a constantly moving landscape of hills and valleys that guide the flow of the river/ consciousness.  Imagine experience as pre-carved tributaries and attention/ prediction influencing the landscape ahead of the flow; pre-charged/ semi-activated networks poised to fire depending on the current input streams. At any point along the flows journey there are multiple valleys of choice… a dry riverbed is a prediction of where the water can/ might flow.  The actual path the water eventually takes is a property of the water not the landscape... top down/ bottom up.

Nice... 

Also, when symbols come into play, they can completely change the game. If you ask somebody to pick up all verbs in the JPG above, a huge process begins in the mind of the one you asked to. It involves moving focus on the image, constructing and maintaining a list (which requires focusing on it quite often), trying to read blurry words, guessing what these words are using their neighborhood, ...etc, all of this because of a few spoken words "pick up all verbs" which have nothing to do with the original perception of this image. I mean, there's no "sensual" relation between the image and the directive.

[keghn]

 I talked about the border that are to take care of the unpredictable object coming in to the video, earlier here above. 
 But there is one special place on the upper border that is a black hole. When the "MAIN FOCUS" move to this it enters 
enters into 3 D simulator space that has three dimension memory, construction memory, and self learned controls of N space.
 Hour can spent in here. To move out the main focus there is little white light at bottom middle. 
 on the side of the video frame bore are the system emotional state markers. 

  After spending hours in 3d simulator space the main focus returns and less than thousandth of a second has past in the real world.

 My AGI will record information in video frames. The first frame will be real data and then next data frames will only the change
between frames. pixel are read and writable.
 Every few thousand frames real data will be recorded. The frame will be mapped out. the first 1000 x 1000
pixel will be left eye. next 1000 x 1000 will be right eye. The next 1000 x 40 pixel will left ear recording. Next 1000 x 40 will
be right ear.
Next 1000 x 40 will be audio output.
 The next 1,000,000 x 1,000,000 will be skin pixels.
 Then one pixel for each joint position. One pixel for each motor for applied motor power value.

 Between each video frames are the predicted frames and a second frame showing the error between predictions. Next
frame we are back to real data frames.

 The logic that deals with prediction and action is hidden away in the dark place somewhere else.

[keghn]

Jürgen Schmidhuber's Solution to AI Consciousness: 

[Art]

Quite an interesting approach but someone get this guy some decaf!

[Zero]

Sounds correct. They made it.

[keghn]

 Now all he has to do is mash a compression graph with a cyclic graph such as this one below.
 
Adaptive Template Model of Intelligence: 

https://signifiedorigins.wordpress.com/2017/09/08/a-template-driven-approach-to-high-level-pattern-recognition/

 There are compression edges into nodes and there are traveling sales man edges into nodes.
 This confusion has stymied almost  all AI scientist.