humanAI.com

 
taskengine
timemachine
patents
notes
donation

 

 

 

                                  << content                              Chapter 6

Robot pathways and encapsulated work (part1) 

In the last section I talked about the clarity tree and how it works.  In this section I will deal with the other two pathway types:  robotís pathways and encapsulated work (or hidden instructions).  I think that the last two pathway types have to be explained simultaneously instead of separately.

The robotís pathways are the data sensed from the robot while operating the plane.  The thoughts of the robot are also stored in the pathway.  The robotís pathway contains the 4 different data:  5 sense objects, hidden objects, activated element objects and pattern objects.  Language is very important to intelligence because it brings order to a chaotic world.  Life is dynamic and no one experiences the same situations twice.  They can experience similar situations, but not the exact situations.

The robot controls the eyes and ears of the plane and makes decisions for the plane to act intelligently in the future.  Commands are given by the robot to the different machines inside the plane to operate.  The robot also identifies objects, set goals, solve conflicts of tasks, avoid obstacles, focus on objects, learn knowledge, apply knowledge, solve problems, give commands to other people and so forth. 

The robot essentially is the brain that controls all aspect of intelligence for the plane.  The clarity tree is there to help the robot understand the environment with greater detail.  The clarity tree also provides data that the robot is and isnít aware off.  For example, the robot is only aware of certain objects in the human visibility level, but it isnít aware of any objects in the atom visibility level. 

 

A team of robots working to control the plane

In Star Trek, there are multiple people that work together to control the plane.  The captain gives the orders and the other people follow the instructions given by the captain.  There might be a hierarchical structure of people working together.  The captain might have a chief engineer that gives orders to lower level workers or a first officer that gives orders to other workers to handle secondary tasks.  

Thus, in addition to the robotís pathways there can be many robots that are working together to operate the plane.  If all these robot pathways self-organize, a station pathway is created.  A station pathway is one universal pathway that contains multiple robot pathways that have relational links to one another.  FIG. 25 is a diagram depicting a station pathway.  There are 5 robots all together.  The main robot is the leader or captain that decides how the plane will operate.  The 4th robot has its own worker that takes orders from only the 4th robot. 

FIG. 25

Station pathways can be structured in any business or organization.  A hierarchical structure of a business can be created and represented by a station pathway.  A school administration system can be created and represented by a station pathway.  Each member of the station pathway knows the rules, the objectives of the team and the powers of the team from common knowledge.  These common knowledge can be found in books, or instruction manuals, etc.  For example, a worker knows his own rules, powers, and objectives from business school.  If the worker is the president of a company, he knows what powers he has and what rules other lower level workers must follow.  The policy from the company will give a more definite guideline to behave in the company.  This guideline should set the environment so that all members of the company know what rules to follow, know their status in the company, and what their objectives are.    

Each member that is in the plane has their own responsibilities and duties.  Each member is also intelligent at a human-level. 

For simplicity purposes letís say that the plane was controlled by only one robot.  All operations of the plane are commanded by a single robot. 

 

Using language to organize data in the clarity tree

Language is the key to establishing more relationships between the clarity tree and the robotís pathways.  The clarity tree has only commonality groups (by default), but it doesnít have any learned groups.  The intelligence from the robot will give the objects in the clarity tree (especially the human visibility level) the ability of language.  The robot will identify objects from activated element objects and these activated element objects serve as the learned groups.  For example, in the human visibility level, if there is a cat that is identified, the pathway from the robot will identify that as the word:  ďcatĒ.  This learned word ďcatĒ identifies what the visual cat is in the clarity tree. 

As stated numerous times in the past, a visual cat can come in different sizes, shapes and color.  The learned word ďcatĒ identifies the visual cat as one fixed word.  A car accident can be presented in infinite ways, but the learned words ďcar accidentĒ identifies that event into a fixed word/s.      

FIG. 26 is a diagram depicting the two relational links between the clarity tree and the robotís pathways.  The human visibility level is referenced because that is the sight the robot sees.  The 5 senses of the robot are referenced and the conscious thoughts of the robot are also referenced.   

FIG. 26

The interesting thing about the relational links between the robotís pathways and the clarity tree is that the clarity tree can reference words/sentences to its lower levels.  FIG. 27 is a diagram depicting the learned words/sentences in the human visibility level are carried over to the molecule visibility level.  Next, the learned words/sentences in the molecule visibility level are carried over to the atom visibility level.  The robotís intelligence provides these learned words/sentences and identify and prioritize visual objects (or any other 5 sense data). 

FIG. 27

 

The encapsulated work for the plane

The universal computer program must be used to encapsulate work for the plane (atom manipulator).  Creating a software to control how a laser system shoot photons at surrounding atoms and to make the atoms behave a certain way is very very difficult.  My first computer program was the universal AI program which trains machines to do tasks with human visibility like drive a car, fly an airplane, mow the lawn, or vacuum the carpet.  Building a machine to do things at an atomic level is infinitely harder. 

The clarity tree is very valuable because work has to be done by different robots on all levels of the tree.  Work must be done at the human visibility level, at the molecule visibility level, and at the atom visibility level.  These work are not done by one robot, but by a hierarchically structured team of robots, each having their own responsibilities and duties. 

Also, work has to be done in fragmented sequences, whereby work is encapsulated in fixed interface functions so these fixed interface functions can be reused in the future.  Think of one control function in the plane as a very long station pathway.  All sections of the station pathway have to be trained, starting from the lower levels and working up towards the top levels.  FIG. 28 is an illustration of one long station pathway to control one function for the plane.  Multiple virtual characters, structured in a hierarchical manner, are working together to make this function work properly.   

FIG. 28

Since the station pathway canít be trained all at once, it is the job of each section of the station pathway to encapsulate their work using the universal computer program.  FIG. 29 shows that each section has to be trained from the bottom first and then trained towards the top levels.  It canít be trained from the top to the bottom because if encap3 was trained first the desired output will be wrong and further because encap3 needs encap2 and encap1. 

However, when all sections of the station pathway are trained adequately, any section or combination of sections can be trained and each trained section will be stored in their respective areas.  For example, if all sections in the station pathway are trained, encap3 or encap2 or encap1 or element combinations from each section can be trained.   

The idea is to separate sections of the station pathways into independent sections.  What sections in the station pathway should be grouped together independently and assigned to a fixed software function?  People can do research and find the best groupings.  These research are then put into books and should be widely read by people who are in the field.  Of course these research methods donít have to be fixed; if other writers find a better method they can also replace the old method with newer methods. 

 

TV monitors to view different levels of the clarity tree 

As stated before, the virtual characters have to do work on many different levels in the clarity tree.  Each virtual character might have to manage multiple visibility levels in order to do work.  The TV monitor is the media that will allow virtual characters to view different visibility levels in the clarity tree.  Software will be included to switch from one level to the next or to view multiple visibility levels at the same time.  For example, there can be two monitors.  One monitor will display human visibility and the other monitor might display molecule visibility. 

A hierarchical team structure is more complex.  Letís say that there is a captain and he is in charge of 2 workers.  The captain is viewing the environment using human visibility and the workers are viewing the environment using molecule visibility.  The workers will do their jobs according to the commands given by the captain, but the captain isnít concerned with the molecule level, he is concerned with the overall human visibility level.    

I will give another example to better illustrate my point.  A captain is viewing the environment using human visibility.  The captain has 1,000 workers that are controlling lasers that will shoot molecules and to force atoms to behave in a certain way.  These workers are also assigned to certain areas of the environment.  The workers are given orders to push atoms in their area toward a targeted location in the environment.  These workers are not aware of how their job will affect the overall job of all workers.  The captainís responsibility is to monitor what happens in the human visibility level and to use software to communicate with the lower level workers and give them instructions so that a desired goal is met.     

FIG. 30 is a diagram depicting a station pathway that is viewing different visibility levels in the clarity tree.  The main virtual character is viewing D1, virtual character2 is viewing D2 and D3, and virtual character3 is viewing D4.  They are working as a group using the data from the clarity tree.  

FIG. 30

Another very interesting note is that as each virtual character does work on the different clarity levels (D1-D4), the virtual character is identifying objects, actions and events.  Sentences and words are assigned to objects/actions/events that are in the clarity tree via the virtual characters conscious thoughts.

Thus, the main robot that is controlling the actual plane is using its conscious to identify objects, actions and events in the human visibility level.  On the other hand, the virtual characters who are working on the other lower visibility levels are also using their conscious to identifying objects, actions and events.  Language in terms of sentences and words bring order to chaos.  It will further help organize the data in the clarity tree.  FIG. 27 shows that learned words/sentences from any level of the clarity will reference upper or lower levels.  Just like how the robot controlling the plane has the ability to assign language to the human visibility level, the virtual robots can also assign language to the lower levels of the clarity tree.  Commonality groups and learned groups will reference each other from different levels in the clarity tree.  

FIG. 27

 

Determining visibility levels in the clarity tree

The signalless technology creates the clarity tree.  It will use the cameras on the plane to form a reasonable clarity tree.  Another factor is the signalless technology will search in memory for any pathway matches to the current pathway.  The pathway matches found in memory will further help to generate an optimal clarity tree.  The pathways in memory self-organize and they are structured in terms of priority Ė the most important objects in a pathway are delineated and the least important objects are not delineated.  By finding the best match in memory the pathway matched will tell the signalless technology which objects in the clarity tree are important and which objects in the clarity tree are not important.  It will also determine how many levels to include in the clarity tree and what these levels are.  

For example, in FIG. 30, D1, D2, D3, D4 are visibility levels that are used by virtual characters.  They will be created based on how important they are to the team work.  Maybe D1 is very important, so the signalless technology creates a detailed pathway for that level.  Maybe D4 is the second most important level, so the signalless technology creates a medium detailed pathway for level D4. 

How many levels the signalless technology will generate for the clarity tree will depend on what information is stored in memory.  If there are lots of pathway matches found in memory, there will be many levels to the clarity tree.  If there are little pathway matches found in memory, there will be small levels to the clarity tree.  Itís the same with human beings and how they learn things.  When we search for a face in memory there are lots of information about faces so our brain have more detailed information about faces.  When we search for fingerprints in memory there are little information about finger prints, so our brain have little information about fingerprints.  Even though faces are very similar to one another we are able to recognize the details to distinguish one person from another.  On the other hand, the fingerprint has little information in memory, and therefore a person canít recognize details on the fingerprint.        

The more information that is stored in memory that matches to the current pathway (the current environment) the more visibility levels the clarity tree will have.

If there are little or no pathway matches found, the signalless technology will defaultly create its own visibility levels.  It will learn from experience to find out which objects, actions or events in the clarity tree are important and it will adapt.  The next time it encounters a similar situation it will know what to include in the clarity tree.  

Other factors also determine how many levels to create in the clarity tree.  Pain and pleasure felt by the virtual characters will prioritize objects.  Which objects causes pain and which objects causes pleasure is very important to determine which objects/actions/events are important.  For example, if a virtual character touches a needle, the pain will cause the virtual character to make the needle have higher priority because the needle caused great pain for the virtual character.  While the virtual characters are working on the different levels of visibility, the pain/pleasure they feel will prioritize objects in the clarity tree. 

Another important thing is where does the focus area begin and end should be based on what parts of the environment are important.  And like I said before, the focus areas should be based on the pathways in memory.  The signalless technology can also have a default focus area or a focus area depending on the cameras visibility.  The signalless technology can also have software programs to create more information to be included in the clarity tree besides the information stored in pathways in memory.  The clarity tree should provide ďextraĒ information that the pathways in memory donít have.       

 

<< content               next chapter >>

 

 

  

Home | Task Engine | Time Machine | Patents | Notes | Donation 

Copyright 2007 (All rights reserved)