|
Software Patent Abstract
A method of creating human level artificial intelligence in machines
and computer software is presented here, as well as methods to simulate
human reasoning, thought and behavior. The present invention serves
as a universal artificial intelligence program that will store,
retrieve, analyze, assimilate, predict the future and modify information
in a manner and fashion which is similar to human beings and which
will provide users with a software application that will serve as
the main intelligence of one or a multitude of computer based programs,
software applications, machines or compilation of machinery.
Software Patent Claims
1. A method of creating human level artificial intelligence in machines
and computer based software applications, the method comprising:
an artificial intelligent computer program repeats itself in a single
for-loop to receive information, calculate an optimal pathway from
memory, and taking action; a storage area to store all data received
by said artificial intelligent program; and a long-term memory used
by said artificial intelligent program.
2. A method of claim 1, wherein said for-loop contain instructions
that said artificial intelligent program must accomplish within
a predefined fixed time limit, for example, 1 millisecond, 10 millisecond,
86 millisecond, instructions in said for-loop comprising the steps
of: entering said for-loop; receiving input from the environment
in a frame by frame format or movie sequence, each frame containing
at least one data comprising of at least one of the following senses:
sight, sound, taste, touch, smell or a combination of senses; searching
for said input in memory and finding the closest matches; calculating
the future pathway of the matches found in memory and determining
the optimal pathway to follow; storing said input in the optimal
pathway and self-organizing said input with the data in a computer
storage area called memory; following the future pathway of the
optimal pathway and exiting said for-loop; and repeating said for-loop
from the beginning.
3. The method of claim 2, wherein searching for information is
based on searching for one pathway in memory, which is referred
to as the optimal pathway, and said artificial intelligent program
will take action by following the optimal pathway's future pathway
4. The method of claim 2, wherein searching for the input in memory,
the input called the current pathway, the method comprising the
steps of: using an image processor to break up said current pathway
into sections of data, called partial data; searching for each of
the partial data in memory using randomly spaced out search points;
each search point will collaborate and communicate their search
results with other search points to converge on the pathways that
best match said current pathway until the entire network is searched.
5. The method of claim 4, wherein each search point will communicate
with other search points on search results with at least one of
the following: successful searches, failed searches, best possible
searches and unlikely possible searches.
6. The method of claim 4, wherein each search point has a priority
number, and determining said priority number comprises of at least
one of these criteria: the more search points that merge into one
search point the higher said priority number; the more matches found
by the search point the higher said priority number; and the more
search points surrounding that search point the higher said priority
number.
7. The method of claim 6, wherein the higher said priority number
the more computer processing time is devoted in that search point
and the lower said priority number the less computer processing
time is devoted in that search point.
8. The method of claim 3, wherein if the search function doesn't
find an exact match in memory said artificial intelligent program
will attempt to fabricate pathways and fabricate future pathways
by using at least one of the four deviation functions: fabricating
pathways using minus layer pathways, fabricating pathways using
similar pathways, fabricating pathways using sections in memory,
and fabricating pathways using the trial and error function.
9. The method of claim 2, wherein calculating the future pathways
comprises: designating a current state in a given pathway and determining
all the future sequences in said pathway; adding all the weights
for each possible future sequences; calculating the total worth
of each possible future pathway and ranking them starting with the
strongest long-term future pathway.
10. The method of claim 1, in which the storage of data is based
on a network contained in a 3-dimensional grid, said data being
represented by objects comprising of at least one of the following:
visual images, sound, taste, touch, smell, math equations, or combination
of objects.
11. The method of claim 10, wherein the 3-dimensional grid stores
at least one data structured tree, each tree can grow or shrink
in size based on the amount of training, and each tree can break
apart into a plurality of sub-tree branches when data is forgotten.
12. The method of claim 10, in which the storage space uses a 3-dimensional
grid to contain all the pathways from input; and each pathway is
subject to space in the 3-dimensional grid where 2 data can not
occupy the same space at the same time.
13. The method of claim 10, wherein during self-organization in
the 3-dimensional grid said artificial intelligent program will
designate a given radius, centered on the input data, to bring common
groups closer together; data outside of said radius will not be
affected while data in said radius will be subject to changes.
14. The method of claim 10, wherein each data comprises two types
of connections with other data in memory and are independent of
each other: sequential connections, which is best represented as
a frame by frame movie; and encapsulated connections which are objects
that are contained in another object, for example, pixels are encapsulated
in images, images are encapsulated in movie sequences, and movie
sequences are encapsulated in other movie sequences.
15. The method of claim 14, in which the sequential connections
are used for predicting the future while the encapsulated connections
are used for storing and retrieving data from memory.
16. The method of claim 2, wherein self-organizing of data, also
known as the rules program, finds association between objects in
memory, the method comprising the steps of: designating an object
from input as a target object; searching and identifying said target
object in memory; designating the objects surrounding said target
object in memory and the objects surrounding said target object
in the input space as the element objects; and bringing the element
objects closer to said target object based on association.
17. The method of claim 16, wherein the association between target
object and the element object further comprising: the more times
the target object and the element object are trained together the
stronger the association; and the closer the timing of the target
object and the element object are the stronger the association.
18. The method of claim 16, in which said artificial intelligent
program will use the rules program to create the human conscious,
the method comprising the steps of: searching and identifying target
objects from input; gather all the closest element objects from
all the target objects found in memory; determining which element
objects will be activated; and activating each of the qualified
element objects in linear order.
19. The method of claim 18, wherein activating element objects
will result in conscious thoughts equivalent to human beings, said
conscious thoughts being represented by instructions, in the form
of language or visual images, that will guide said artificial intelligent
program to execute at least one of the following: solve arbitrary
problems, provide meaning to language, give information about an
object, and provide general knowledge about a situation.
20. The method of claim 16, wherein meaning of objects, most notably
meaning to language, occurs when two or more objects fall within
the same assign threshold, for example, a sound of cat, the visual
text cat, and the visual floater of cat are stationed in the same
assign threshold, therefore all three objects have the same meaning.
21. The method of claim 16, wherein self-organization of data comprises
two types of groups: learned groups; and commonality groups.
22. The method of claim 21, wherein said commonality group is represented
by any 5 sense traits or hidden data that two or more objects have
in common such as common traits represented by sight, sound, taste,
touch, smell or hidden data set up by the programmer within these
5 senses.
23. The method of claim 21, wherein said learned group is represented
by two or more objects that have strong association to one another;
particularly two or more objects that are stationed in the same
assign threshold.
24. The method in claim 10, wherein the 3-dimensional storage grid
uses the 2-dimensional movie frames and store them in such a way
that said 2-dimensional movie frames produces a 3-dimensional environment.
25. A method to mimic long-term memory similar to human beings
in claim 2, the method comprising: a timeline, with increments of
1 millisecond, that contain reference points to the time movie sequences
occurred; said timeline has reference pointers to movie sequences
stored in memory; and said artificial intelligent program uses said
timeline to find patterns to intelligence and conscious thought.
26. A method to create an N-dimensional object from 2-dimensional
sequential movie frames, said N-dimensional being represented as
any-dimensional, the method comprising the steps of: using an image
processor to delineate moving or non-moving image layers from one
frame to the next in said 2-dimensional movie; using the self-organization
technique in said artificial intelligent program to find repeated
patterns based on colored pixels from frame to frame; determining
what image layers belong sequentially from frame to frame and designating
the strongest sequential image layers as the center of said N-dimensional
object; and determining a predefined range of how fuzzy said N-dimensional
object can be and anything that falls within this fuzzy range will
be considered said N-dimensional object.
27. A method of claim 4, wherein said current pathway comprises
at least one of the following data types: 5 sense data or commonality
groups; activated element objects or learned groups; hidden data
and; patterns;
28. A method of claim 27, wherein each data type have their own
encapsulated format.
29. A method of claim 27, in which said hidden data are created
during runtime based on the 5 sense data, said hidden data for a
visual object comprises: a normalization point of said visual object;
an overall pixel count of said visual object; a scaling analysis
of said visual object, a rotation analysis of said visual object,
a movement path of said visual object, a movement distance of said
visual object, a number of changes of a movement direction of said
visual object, and a number of contacts between said visual object
and other visual objects.
30. A method of claim 27, wherein said patterns uses internal functions
to assign instructions in pathways to extract data from memory and
predict the future.
31. A method of claim 30, wherein said internal functions include:
the assignment statement, searching for data in memory, determining
the distance between data in the 3-d environment, rewinding and
fast-forwarding in long term memory to get data, and determining
the strength of data in memory.
32. A method of claim 30, wherein said artificial intelligence
program compares data from similar pathways in memory to find said
patterns
33. A method of claim 10, wherein if there are multiple copies
of an object in memory each copy of said object will have a reference
pointer to a masternode, said masternode being represented as the
copy of said object with the highest powerpoints
34. A method of claim 33, wherein training of an object occur in
a global fashion where all copies of said object's powerpoints will
be modified, the method comprising the steps of: said object sends
a signal to the masternode to identify itself and; said masternode
will modify most copies of said object in which the stronger the
pointer connection the stronger the modification.
35. A method of claim 10, wherein the priority of objects in a
given pathway state is determined by at least one of the following
factors: said artificial intelligence program uses pain and pleasure
in which said artificial intelligence program identifies objects
that causes the pain or pleasure and; said artificial intelligence
program compares data in similar pathways to determine wither or
not an object causes the pathway to change its future course.
36. A method of claim 18, in which the steps to extract element
objects from a target object comprises: said target object sends
a signal to the masternode to identify itself and; said masternode
will extract element objects from all copies of said target object
based on the connection pointers, wherein the stronger the connection
pointer the higher the priority of the element object.
37. A means by an artificial intelligence program to use language
in a fuzzy logic manner to accomplish at least one of the following
functions: storing and organizing 5 sense data in a computer readable
memory or network; predicting the future without the aid of heuristic
search algorithms, discrete mathematics, language parsers, planning
programs, genetic programming, and probability theories; predicting
the future with the aid of heuristic search algorithms, discrete
mathematics, language parsers, planning programs, genetic programming,
and probability theories; planning tasks and solving interruption
of tasks; defining the rules of an image processor to extract information
from pictures or movie sequences and; creating logic and reasoning
from 5 sense data;
Software Patent Description
CROSS REFERENCE TO RELATED APPLICATIONS
[0001] This is a Continuation-in-Part application of U.S. Ser.
No. 11/744,767, filed on May 4, 2007, entitled: Human Level Artificial
Intelligence Software Application for Machine & Computer Based
Program Function, which claims the benefit of U.S. Provisional Application
No. 60/909,437, filed on Mar. 31, 2007.
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
[0002] (Not applicable)
BACKGROUND OF THE INVENTION
[0003] 1. Field of the Invention
[0004] This invention relates generally to the field of artificial
intelligence. Moreover it pertains specifically to human level artificial
intelligence for machines and computer based software.
[0005] 2. Description of Related Art
[0006] For 60 years, ever since artificial intelligence has been
around, scientists have long to build a machine that can think,
reason, behave, and act like a human being. The problem with current
AI software is that they cater to parts of human intelligence and
not human intelligence as a whole. This is why there are so many
subject matters related to artificial intelligence.
[0007] One aspect is the fact that no one has defined what the
conscious is? The conscious is highly debated by both psychologists
and AI researchers. In order to build a human brain the conscious
must be defined. This would include: what the conscious is, how
does the conscious work, and what are the computer codes to implement
the software to a conscious?
[0008] Building a network that will store, retrieve, and modify
information is another aspect that must be considered. The internal
data in neurons and how the dendrites work has baffled many AI researchers.
Neural networks try to resemble how neurons work but there are many
unanswered questions with those AI programs and they don't work
very well. The growing problem of how does the data get stored in
memory and how does the data get retrieved by the host is still
a mystery. What exactly are the data stored in the neurons is also
something that has never been explained.
[0009] Another aspect is the field of reasoning and probability
in machines. Currently, Bayesians probability theories, semantic
networks, discrete mathematics, and language parsers are used in
combination to produce a machine that can learn language/knowledge
in a limited environment. The idea was to build something that can
learn and understand language and to use the language to make the
machines learn things from its environment. However, this is complicated
by the fact that it is very difficult to build a machine that can
learn language using the current AI methods. Even language that
a 5 year old is capable of learning is very difficult to do in machines.
SUMMARY OF THE INVENTION
[0010] To solve the mentioned problems above, the present invention
proposes a totally different way of building a human robot. This
would include defining/building a conscious, building a network
to store/retrieve/modify large amounts of information, building
a machine that can learn language and common sense knowledge, and
building a machine that can learn probability and reasoning. In
addition to this, the invention not only has the capability of human
intelligence but the capability to acquire intelligence that "exceeds"
human intelligence.
[0011] There are thousands of ways of building a human brain. This
human level artificial intelligence program is a collection of 6
years of designing and implementing a software that I think will
produce human intelligence. The HLAI program is a computer brain
that can predict the future. The AI software can be applied to all
machines and the machine will behave intelligently at or similar
to human intelligence. If the human level AI is applied to a car
then the car will drive by itself from one location to the next
in the safest and quickest way possible. If the HLAI is applied
to a plane then the plane will fly by itself from one place to the
next in the safest and quickest way possible. If the HLAI is applied
to a videogame then the AI can play any game for that videogame
system. Just like humans, the AI program uses knowledge from the
past to predict what will eventually happen in the future. By giving
the AI the ability to see into the future it can anticipate what
will eventually happen next and take the best course of action.
[0012] A camera is used to interface the HLAI program with all
the different machines. The program will store all the frame-by-frame
video in memory in an organized way. My program can store large
amounts (almost infinite) hours of video in memory and the retrieval
program will get the video clips in a quick way using multiple search
points. This is revolutionary because it would mean that the computer
will never run out of disk space (current neural networks can't
do this). The program also self-organizes all the data in memory
so that common video clips will be stored in the same area. The
storage part of the program works by storing each frame of the movie
in a 3-d environment. The result is the 3-d representation of all
the movies. The 3-D environment is actually the average of all the
movies stored in memory. Theoretically, this is how humans store
information in memory
[0013] The idea behind the memory of the AI is to store the most
important pathways (movie sequences) and to forget the least important
pathways. The network uses strength of node/s to represent any repeated
data. The more a pathway is trained the stronger the node/s become.
The less training it goes through the less strength the node has.
The length of the pathway also grows with more training and the
length of the pathway shrinks with less training.
[0014] The present invention is novel because it solves 80 percent
of all problems facing the field of artificial intelligence. Some
of the features that are novel in the present invention are: [0015]
A. The AI can learn common sense knowledge and language without
language parsers, discrete mathematics, semantic networks, probability
theories, or any type of modern day AI technique/s. [0016] B. The
AI is capable of learning what is known as universal language. Instead
of limiting the language to English the AI can learn Chinese, German,
Arabic, Korean, Dutch, Spanish, French or any language, even alien
language. [0017] C. It can store large, "almost infinite",
amounts of video or pictures and the data can be retrieved quickly.
[0018] D. In prior art, storing all possible outcomes of a 2-player
game in memory is impossible. The total possible outcome of a chess
program is 10 to the 40.sup.th power and the total combinations
of the outcome are infinite. My program can store all the possible
outcomes of a chess program (which amounts to infinite data). A
more complex form of the chess program is movie sequences from real
life or videogames. My program can store the total possible outcomes
of movie sequences as well. [0019] E. In prior art, the majority
of 2-player AI games such as chess, and checkers use expert systems
to calculate future steps during runtime. My program stores all
the possibilities in memory and uses the stored data to predict
the future (given that a 100 percent pathway match is found in memory).
My program uses fuzzy logic to predict the future for similar or
non-existing pathways in memory. [0020] F. There is no need to insert
rules into the network because the rules are learned through training.
If you apply this program to a car, all the rules of driving are
learned by observation. An expert trainer has to drive the car and
the AI must observe, store and average all the training data in
memory. When the data is averaged out the AI will understand the
rules of driving. [0021] G. The method the AI uses to retrieve information
is faster than any search algorithm in computer science. The timing
of the search is considerable lessened as more data gets inserted
into the network. [0022] H. No modern day AI technique is used to
learn probability and reasoning. The AI learns probability and reasoning
through patterns. I set up the different patterns in the system
and the AI finds these patterns. [0023] I. The HLAI program is versatile
and can be applied to all machines including: cars, trucks, buses,
planes, forklifts, computers, human robots, houses, lawnmowers,
radios, phones, and even toaster ovens. "All" machines
can be hooked up to the HLAI and that machine will act intelligently
at or above human intelligence. [0024] J. The HLAI has no boundaries
as to its application. It not only is a revolutionary technology
applied to computer science, but other disciplinary fields such
as biotechnology, engineering, aero dynamics, chemistry, medicine,
genetic engineering, and mathematics. The novel things that can
be created from this invention are: a software that can predict
an earthquake or hurricane one year in advance, a humanoid robot,
a machine that can predict the future and the past with pinpoint
accuracy, automated software to do all human jobs including: driving,
surgery, retail, technical tasks, operating cameras for movies and
tv, hair cuts, make-up, construction, building houses, fighting
a war and so forth. Anything that a human or a group of humans can
do this invention will also be able to do.
[0025] This patent is very long, 206 pages including drawings.
I feel the need to disclose all information about this invention
in a complete and concise manner so that the reader will have a
better understanding of how "human intelligence" is reproduced
in a computer. The outline of the patent is done in a computer science
manner where the inventor discusses the basic functions of the AI
program first and then dives deeper and deeper into the details.
The inventor tries to introduce information in linear order. However,
some information are repeated or revisited in certain parts of the
patent.
BRIEF DESCRIPTION OF THE DRAWINGS
[0026] For a more complete understanding of the present invention
and for further advantages thereof, reference is now made to the
following Description of the Preferred Embodiments taken in conjunction
with the accompanying Drawings in which:
[0027] FIG. 1 is a software diagram illustrating a program for
human level artificial intelligence according to an embodiment of
the present invention.
[0028] FIG. 2 is the software diagram of the present human level
artificial intelligence program presented in a different way.
[0029] FIG. 3 is a diagram depicting self-organization of data
in memory.
[0030] FIG. 4 is a diagram depicting the current pathway during
each iteration of the for-loop in FIG. 1.
[0031] FIG. 5 is a diagram demonstrating how conscious thoughts
are used to interpret grammar.
[0032] FIG. 6 is a diagram depicting the data structure of memory.
[0033] FIG. 7 is a flow diagram depicting the searching of data
from FIG. 6.
[0034] FIG. 8 illustrates the search process.
[0035] FIG. 9 is a diagram to illustrate the searching process
using both commonality groups and learned groups.
[0036] FIGS. 10-11B are diagrams demonstrating sequential connections
and encapsulated connections.
[0037] FIG. 12 is a diagram of 2-d data structured trees representing
conventional networks, hashtables, vectors, or linklists.
[0038] FIG. 13 is a diagram of 3-d data structure for the present
invention.
[0039] FIG. 14 are diagrams showing the weights of sequential connections
and encapsulated connections.
[0040] FIGS. 15A-15C are diagrams depicting the rules program.
[0041] FIG. 16 is a diagram to demonstrate how the rules program
assigns meaning to sentences.
[0042] FIGS. 17-18 are illustrations to demonstrate image layers.
[0043] FIGS. 19-20 are illustrations to demonstrate how the rules
program assign meaning to nouns and verbs.
[0044] FIGS. 21A-21B are diagrams to illustrate how the mind produces
conscious thoughts.
[0045] FIGS. 22-24 are illustrations to demonstrate the 4 deviation
functions.
[0046] FIGS. 25-27C are diagrams illustrating examples of how the
present invention can demonstrate human intelligence.
[0047] FIGS. 28A-28D are diagrams to illustrate how pathways in
memory can form complex intelligence.
[0048] FIGS. 29-33 are diagrams to demonstrate how the AI program
creates templates and how the templates are trained in memory.
[0049] FIG. 34 are diagrams to demonstrate how templates are used
to lengthen pathways in memory.
[0050] FIGS. 35A-35D are diagrams illustrating the process in FIG.
34.
[0051] FIG. 36 is a flow diagram depicting the process of how objects
are trained in memory.
[0052] FIG. 37 is a diagram depicting the structure of repeated
objects in memory.
[0053] FIG. 38 are diagrams depicting the rules program.
[0054] FIGS. 39A-39B are diagrams illustrating the process of extracting
element objects from a target object and activating said strongest
element objects in linear order.
[0055] FIGS. 40A-40B are diagrams depicting human thoughts.
[0056] FIGS. 41A-41D are different examples of the ABC block problem.
[0057] FIG. 42 is a diagram illustrating grouping of encapsulated
data between hidden objects.
[0058] FIG. 43 is a diagram showing the different times events
occur.
[0059] FIG. 44 is a diagram showing decision making by the AI program.
[0060] FIGS. 45A-45D are illustrations showing how learned groups
and commonality groups organizes face images.
[0061] FIGS. 46A-46F are illustrations demonstrating how moving
objects self-organizes in memory.
[0062] FIGS. 47A-47B are flow diagrams depicting the process of
how newly created objects are trained in memory.
[0063] FIGS. 48A-48B are diagrams demonstrating the 2 types of
data in the current pathway: 5 sense data and activated element
objects.
[0064] FIG. 49 is a flow diagram depicting a hidden object or meaning
forget information.
[0065] FIG. 50 is a diagram depicting how the AI program matches
pathways in memory.
[0066] FIG. 51 is a diagram depicting a target object and its activated
meaning.
[0067] FIGS. 52-53 are diagrams depicting how the AI program matches
pathways in memory.
[0068] FIGS. 54A-54B are illustrations demonstrating how the ABC
block problem self-organizes in memory.
[0069] FIG. 55 is a diagram depicting the organization of data
in memory based on learned language.
[0070] FIGS. 56A-56B are diagrams demonstrating the 3 types of
data in the current pathway: 5 sense data, activated element objects
and hidden data.
[0071] FIGS. 57A-57B are flow diagrams illustrating how commonality
groups or 5 sense data forget information.
[0072] FIGS. 58A-58D are diagrams illustrating how learned groups
or activated element objects forget information.
[0073] FIG. 59 is a flow diagram illustrating how hidden data forget
information.
[0074] FIG. 60 is a flow diagram further illustrating how learned
groups or activated element objects forget information.
[0075] FIGS. 61A-61B are diagrams illustrating how the AI program
reads in the word bat.
[0076] FIG. 62 is a diagram depicting multiple learned groups assigned
to a cat floater.
[0077] FIGS. 63A-63B are diagrams demonstrating the 4 types of
data in the current pathway: 5 sense data, activated element objects,
hidden data and patterns.
[0078] FIGS. 64A-64C are flow diagrams showing how the AI program
finds patterns to similar pathways and output a universal pathway.
[0079] FIGS. 65A-65B are diagrams depicting how the AI program
assigns hierarchical groups as variables in a universal pathway.
[0080] FIG. 66 is a diagram showing the different times events
occur.
[0081] FIG. 67 is an illustration of visual text words in 3-d space.
[0082] FIG. 68 is an illustration of a mouse and the text word
mouse.
[0083] FIG. 69 is a diagram of conscious thought when the AI program
encounters the word mouse.
[0084] FIG. 70 is an illustration of how the AI program identifies
the word mouse in the movie sequences.
[0085] FIGS. 71A-71B is an illustration of how the AI program assigns
the word jump to a movie sequence.
[0086] FIG. 72 is a diagram of different sentences assigned to
the same meaning.
[0087] FIGS. 73A-73E are diagrams depicting the process of assigning
different sentences to the same meaning.
[0088] FIG. 74 is a diagram showing the steps of reading in and
interpreting a sentence.
[0089] FIG. 75 is a diagram showing how the assignment statement
is assigned to a sentence.
[0090] FIGS. 76A-76B are diagrams depicting how different sentences
can be interpreted in a fuzzy logic manner.
[0091] FIGS. 77A-77B are flow diagrams showing the different patterns
in a pathway to predict the future.
[0092] FIGS. 78A-79B are diagrams showing internal function: finding
data from the 3-d environment.
[0093] FIGS. 80A-80B are diagrams showing internal function: rewinding
and fast-forwarding in long term memory to get information.
[0094] FIGS. 81A-81B are diagrams showing two internal functions:
finding data from the 3-d environment and rewinding and fast-forwarding
in long term memory to get information.
[0095] FIGS. 82A-82B are diagrams showing a universal pathway of
FIGS. 81A-81B.
[0096] FIG. 83 is a diagram depicting a universal pathway of FIG.
75.
[0097] FIGS. 84-85 are diagrams depicting target objects and activated
element objects.
[0098] FIGS. 86A-86B are diagrams showing sequential sentence association.
[0099] FIGS. 87A-87D are diagrams showing an example of logic and
reasoning.
[0100] FIGS. 88A-88B are diagrams showing an example of an addition
problem.
[0101] FIGS. 89A-89B are diagrams showing an example of an addition
problem similar to FIGS. 88A-88B.
[0102] FIG. 90 is a diagram showing the different times events
occur.
[0103] FIG. 91 is a diagram showing hierarchical learned groups
of numbers.
[0104] FIG. 92 is a diagram depicting the rules program assigning
a word to a meaning.
[0105] FIG. 93 is a diagram depicting numbers being represented
by learned groups.
[0106] FIG. 94 is an illustration showing visual images assigned
to a word.
[0107] FIG. 95 is an illustration showing visual images assigned
to a word similar to FIG. 94.
[0108] FIG. 96 is an illustration showing a diagram of a hierarchy
tree of mammals.
[0109] FIG. 97 is a diagram showing a variance of FIG. 96.
[0110] FIG. 98 is an illustration showing a diagram of a hierarchy
tree of a family.
[0111] FIGS. 99A-99B are diagrams showing how robots can learn
knowledge by observing a situation.
[0112] FIG. 100 is a diagram showing three pathways with their
powerpoints.
[0113] FIGS. 101-102 are diagrams of pathways at different states.
[0114] FIGS. 103A-103D are diagrams depicting logic and reasoning.
[0115] FIGS. 104-107 are diagrams showing the process of planning
tasks and managing interrupted tasks via language.
[0116] FIG. 108 is a diagram showing how the robot reads and interpret
words in a book.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0117] The Human Level Artificial Intelligence program acts like
a human brain because it stores, retrieve, and modify information
similar to human beings. The function of the HLAI is to predict
the future using data from memory. For example, human beings can
answer questions because they can predict the future. They can anticipate
what will eventually happen during an event based on events they
learned in the past.
[0118] There are multiple parts to the program:
[0119] A. storage of data
[0120] B. retrieval of data
[0121] C. the rules program (or self-organization of data)
[0122] D. future prediction
[0123] All these parts of the program work together to produce
the intelligence of the machine. I will outline each of the parts
individually and try to link them together. The next several paragraph
explains how all the parts work together to form the intelligence
of the machine.
[0124] The present invention provides a method of creating human
level artificial intelligence in machines and computer based software
applications, comprising: the AI program repeats itself in a single
for-loop to receive information, calculate an optimal pathway from
memory, and taking action.
[0125] FIG. 1 is a software diagram illustrating a program for
human level artificial intelligence according to an embodiment of
the present invention. First, the AI will get input (current pathway)
from the environment (Step 2). Next, the AI uses the search function
to find the optimal pathway from memory (Step 4). The optimal pathway
is based on two criteria: the best pathway matches 6 and best future
predictions 8. The input data (current pathway) will be stored in
the optimal pathway. The rules program, the self-organizing of data
and the pattern finding are all done at the time the data is stored
in memory. When all the data is stored, the AI will follow the future
pathway of the optimal pathway (Step 10). Finally, the program repeats
itself from the beginning (Step 12).
[0126] The length of the input will be defined by the programmer.
In FIG. 4 the length of the input, or the current pathway, is 3
frames. During each iteration of the for-loop the AI receives one
extra frame from a camera and this frame will be attached to the
front of the current pathway; designated as the current state. The
last frame of the current pathway will be deleted. The current pathway
will be the fixed pathway searched in memory at each iteration of
the for-loop.
[0127] Storage
[0128] Human beings store information in terms of a movie. If that
person lives for 10 years then the brain has to store 10 years worth
of video. If that person lives for 1 thousand years then the brain
has to store 1 thousand years of video. The purpose of the storage
is to collect large amounts of movies and store them in a way that
will minimize repeated data and prevent memory overload. The current
neural networks or compression programs can't do this. My HLAI can
store large amounts of movies in a network where all the data are
interconnected.
[0129] Data is stored in terms of a movie--frame by frame. The
things that can be stored in the frames can range from images to
sound to other senses such as taste, touch, and smell. I call these
data, objects, because they can be "anything". An object
can be a dog barking or a blue pencil or a letter. Objects can also
be encapsulated such as a hand is one object that is encapsulated
in another object, the arm. Objects can also be combined. One example
is the sound of a car zooming by and the images of the car moving.
(When I mention words such as: pathways, data, information, and
movie sequences I'm referring to objects)
[0130] For each data in memory there are two types of connections:
sequential connections and encapsulated connections. Both types
of connections are independent of one another but are used to connect
data in the same storage space. The sequential connections 18 are
shown in (FIG. 10), where each arrow represents a sequential connection.
Data are stored in the frames and the data can be anything. On the
bottom (FIG. 11B) is a diagram of encapsulated connections 22. These
are connection points that states that one object (data) is encapsulated
in another object (data). The AI will be using the sequential connections
to predict the future and the AI will be using the encapsulated
connections for storing and retrieving information from the network
(FIG. 14).
[0131] As the AI learns knowledge from the environment the weights
of the connections (for both connection types) will get stronger
and stronger. In some cases the connections get weaker and weaker
based on external factors such as pain or pleasure. When data is
repeated the data gets stronger. When data is unique and new it
is created. As time passes the data that aren't trained often will
be deleted from the network and data that are trained often is kept
in the network. This is similar to how humans remember things. The
most important information is kept in memory while the minor information
is deleted.
[0132] Data in memory are also organized into two groups: commonality
groups and learned groups. The commonality groups are the groups
that have some form of common physical trait. A man and a women
have common traits. Although they are different they both have 2
arms, two legs, and 1 head. The learned groups are groups that are
learned to be the same. For example, a horse and a pig look absolutely
different. However, they are both animals. The word animal is the
learned group for both the horse and the pig.
[0133] Both the learned groups and the commonality groups must
co-exist in the same storage space. All the data are also encapsulated
within these two groups. In memory, anything that has similar traits
to each other will be grouped and brought closer together. This
is how the data in the network are interconnected and each data
is connected to other data in the network globally. An example of
this is from the diagram (FIG. 6). This diagram displays the level
of encapsulation for visual images and movies. The lowest level
will be the pixels. The pixels are encapsulated in the images. Next,
the images are encapsulated in the frames. Finally, the frames are
encapsulated in the movies.
[0134] In the current neural networks, when data is inserted into
memory, every data in the network must be modified. This can waste
a lot of disk space and computer processing time. The HLAI program
on the other hand only changes specific data in memory but at the
same time preserve the fact that the network is interconnected.
The secret is that when the AI stores a pathway in memory it looks
at its neighbors to find if there are any commonality and learned
groups nearby. When and if the AI finds common groups it will bring
those data with the same group closer together. Referring to FIG.
3, if two identical nodes are close enough (radius 14) they will
merge into one and this will free up disk space. The new nodes will
be created and connected to existing nodes in the network.
[0135] In terms of the topology of storage, data will be contained
in a 3-dimensional grid where the movie pathways are stored as trees
or branches of trees. In FIG. 12 the conventional way of building
trees, networks, hash tables, vector arrays, or linklists will not
work. Most of the data structures used today store information in
one fixed tree with one fixed starting point. This would mean that
in order to store information the tree has to be traversed from
a fixed point and stored in its appropriate area. In FIG. 12 the
relationship between elements A B C in the first tree will not have
any relationship to A B C on the second tree and can not be brought
closer together.
[0136] In a 3-dimensional grid the trees do not have a fixed point
to start from nor does it require traversing the tree to store information
(FIG. 13). The data is not stored in one tree but multiple trees
that grow in size and length. Data in memory can shrink because
data can be forgotten or it can grow if new data is inserted. Sections
of long trees can be broken up into sub trees or it can migrate
from one part of memory to another part of memory (this process
is slow because the network needs time to adequately self-organize
data and to preserve the global data connections).
[0137] One advantage of 3-d storage is that the AI can store pathways
anywhere in the 3-d space without having to search and identify
items from a fixed point. All the trees and all the branches of
the trees can be easily retrieved by the search algorithm discussed
below.
[0138] Another advantage of 3-d storage is that the AI can bring
branches of trees together without traversing the branches of the
trees. In FIG. 13 the AI will bring commonality traits of all the
branches of the trees that fall within a given radius 24. A,B,C
are the common traits. Any data that is contained in the radius
will be subject to self-organization, while data outside of the
radius will not be affected. This will bring relations between data
closer together where the data can self-organize itself only in
specific areas. This will also preserve the fact that all data in
the network are interconnected in a global manner.
[0139] The movie pathways are stored and arranged in memory based
on their sequences. This will create a 3-d environment using the
2-d movie frames. Although the movie will have many variations,
many temporary objects, and many object layers, the function of
self-organization will knit all the data in memory together. Anything
that is stationary is more likely to have a permanent place in memory,
while objects that move a lot is temporarily stored. After averaging
out all the data, the 3-d environment will be established first
because the majority of our environment stays the same. Things like
pedestrians, moving cars, and non-stationary objects are forgotten.
The 3-d environment is considered one big floater because it has
a fuzzy range of itself--the environment can be day or night or
rainy or damaged and so forth but because it falls within the floater's
fuzzy range it will still be identified as the environment (floaters
will be discussed shortly).
[0140] Retrieval of Data in the Network
[0141] The purpose of retrieving data from memory is to find one
pathway, the optimal pathway, that best matches the current pathway.
[0142] For retrieving data from memory, the strength of each data's
encapsulated connections in memory has already been established
based on training (FIG. 14). Searching for data is accomplished
by following the strongest encapsulated connections. This means
that if the AI receives partial data of an image it will follow
the strongest encapsulated connections to get the full data of an
image.
[0143] Retrieving data from the network will require multiple search
points. The AI will randomly pick out search points in the network.
These search points will communicate with each other during the
search process to find the data that it is looking for. This form
of searching for information is faster than any search algorithm
in computer science because it uses multiple search points along
with a form of fuzzy logic to get information. This searching of
data is kind of like throwing ants randomly in a room. At the center
of the room is a piece of candy. As the ants searches for the candy
they will communicate with each other to find the candy. When one
ant finds the candy all the other ants know where the candy is located.
[0144] Each search point will communicate with other search points
on search results such as successful searches, failed searches,
best possible searches and unlikely possible searches. Each search
point has a priority number, and determining the priority number
depend on these criteria: the more search points that merge into
one search point the higher the number, the more matches found by
the search point the higher the number, and the more search points
surrounding that search point the higher the number. The higher
the priority number the more computer processing time is devoted
in that search point and the lower the priority number the less
computer processing time is devoted in that search point.
[0145] The retrieval of data uses both the commonality groups along
with the learned groups to find information. The learned groups
use the top-down search method and the commonality groups use the
bottom-up search method. Both the bottom-up search method and the
top-down search method will be used to search for information. In
(FIG. 7) the search is done using commonality groups. In (FIG. 9)
the search is done with both commonality and learned groups.
[0146] First, the AI breaks up the current pathway into sections.
The current pathway is the pathway the AI is currently experiencing.
The image processor will guide the process of breaking up the data
into sections. Each section will be searched in memory based on
randomly spaced out search points. All searches are done by traveling
on the strongest encapsulated connections. Each search point will
communicate with other search points on possible good searches or
failed searches. The search points will merge together when they
have the same search results and their priority number will be combined.
The better the search result the more search points will be in that
area. This will happen throughout all the search points until they
converge on a (pathway 16) match for the current pathway (FIG. 7).
If the current pathway isn't found in memory the AI will find the
closest match.
[0147] The learned groups are used in the search process to find
data even faster because they can tell the search points what are
continuous frames and what aren't continuous frames. For example,
if a search point finds one cat image in memory then the image sequence
of the cat is also found in memory because visual images are stored
in a 3D environment. In FIG. 9 the X marks the individual search
points. These search points are known as partial data. The purpose
of the search points is to find the whole data. Each search point
will follow the strongest encapsulated connected nodes to find better
matches. Once the whole data is found it will tap into the whole
data's learned group. In this example "A" represent horse,
"B" represent the sun, and "C" represent a tree.
The whole data is the visual image of the horse. Partial data is
the visual head of the horse. When the whole image of horse is found,
that image has a learned group, the word "horse". Once
the learned group "horse" is identified then all the sequential
images of horse from the current pathway will also be identified.
This process will repeat itself for A, B and C. The search points
will keep trying to find better and better matches until the entire
network is searched.
[0148] When the AI locates the optimal pathway (or the best pathway
match) in memory that is where the current pathway will be stored.
But before that can happen a process of breaking down the current
pathway into its encapsulated format must be done. This process
consumes a lot of disk space but is necessary to preserve the global
network. In (FIGS. 11A and 11B) the AI breaks down the current pathway
into its encapsulated format based on the pathways the search function
took to find the optimal pathway 20. This means that pathways that
lead to the optimal pathway 20 are used to break down the input
data into its encapsulated parts. Once the encapsulated format is
created for the current pathway, new data will be created and stored
in its respective area while data already in the network will be
strengthened.
[0149] In (FIG. 11B) the current pathway is broken up into objects
A, B, and C. Then it further breaks down the objects into its encapsulated
objects. Things that make up that object, most notable the strongest
objects, will be broken down. This process will go on and on until
the individual pixels. If this takes up too much disk space and
computer processing time the programmer can define how far the AI
can break down the images. For example, break down images until
the pixels are made up of groups of 6.
[0150] However, understand that the data in memory forgets. Several
hours after the new data is inserted into memory, half of the data
will be forgotten. If the data is trained many times it will stay
in memory permanently, while data that happens coincidentally will
stay in memory temporarily.
[0151] The Rules Program
[0152] Objects
[0153] Objects can be anything. It can be sound, it can be vision,
it can be touch, and so forth. A visual word can be an object, a
sound of a word can be an object, or the visual meaning of the word
can be an object. For different senses the objects can be represented
differently. There is also the consideration of combinations of
objects together such as a visual object in conjunction with a sound
object. A car zooming by is a combination of a visual object and
the zoom sound is the sound object. Or dropping a pencil on the
ground is a combination of visual and sound objects.
[0154] Another factor is that objects can be encapsulated. For
example, a hand is an object that is encapsulated in another object,
a human being. Another example is a foot is an object encapsulated
in another object, a leg.
[0155] The way the program learns these objects is by repetition
and patterns. Each object is represented by strength and if it ever
repeats itself the strength gets stronger. If the object don't repeat
itself then it will forget and memory won't have a trace. 1-d, 2-d,
3-d, 4-d, and N-d objects can be created by repetition and patterns.
[0156] Object Association is the Key to the Conscious
[0157] For each object the AI has to find other objects in memory
that have association. "The more times two objects are trained
together" and "the closer the timing of the two objects
are" the more association the two objects have with one another.
FIGS. 15A-15B are diagrams depicting the rules program. The object
that will be used to find associations is called the target object
26 and the objects that have associations are called element objects
30.
[0158] When the AI recognizes the target object from the environment
it will activate closest element objects that have association to
the target object. There are three types of element objects:
[0159] A. equals (same meaning)
[0160] B. stereotypes
[0161] C. trees
[0162] Equals
[0163] Objects that are very close to each other are considered
"equal". Referring to FIG. 15A-15B, when any element object
34 passes the assign threshold 32 the element object 34 and the
target object 26 are considered equal--they have the same meaning.
(FIG. 15C) One example of this is the sound "horse", if
the sound "horse" is the target object 38 and the element
object 42 that passes the assign threshold is a visual image of
a horse then both the sound "horse" and the visual image
of horse is considered the same.
[0164] Stereotypes
[0165] Stereotypes are facts about the target object. Objects that
are associated with the target object but are not consistent are
stereotypes. These objects are also farther away from the target
object. We look at the fixed object as a part of the overall object.
If the target object is "cat" and "cat" is a
part of "cats don't like dogs", then we can safely say
that "cats don't like dogs" is a stereotype of "cat".
[0166] Trees
[0167] Trees are objects that are usually farther away from the
target object. Sometimes trees have relations to the target object.
A tree is just instructions that people teach you at certain situations.
Timing of the object is the key difference between stereotypes and
trees. This is the most important trait in my program to convey
intelligence. One example of trees is when you cross the street,
the tree "look left, look right and check to make sure there
are no cars before crossing the street" pops up in your mind.
[0168] To better understand about the rules program I will explain
how the HLAI learns language.
[0169] How Human Robots Interpret Language
[0170] When dealing with language there are many AI software that
tries to represent language. Among the most popular categories are:
language parsers, discrete mathematics, and semantic models. None
of these fields (or a combination of them) can produce a machine
that can fully understand language similar to human beings. Designing
a machine that can learn language requires a lot of imagination
and creativity. My design of how to represent language comes from
two sources: Animation and videogames. Mostly videogames because
that is where my key ideas come from.
[0171] Common sense knowledge using language is very hard to represent
on a computer because it's "all or nothing". Either the
computer can understand the language similar to human beings or
they don't understand the language at all. People who clean rooms
for a living not only need knowledge about cleaning rooms but also
common knowledge that humans have. Basic things like: if you drop
something it falls to the ground, if you break the law you will
go to jail, if you throw an egg it will fall and break, if you don't
eat you will get hungry. These are basic knowledge that every human
should know. Machines on the other hand has to be fed the knowledge
manually, unless someone builds a learning machine similar to a
human brain. Even universal learning programs like the neural network
require programmers to manually feed the rules and data in order
for it to work. Like I said it's "all or nothing".
[0172] If there exist a robot janitor and the function of the robot
janitor is to clean the house, what happens when it's mowing the
lawn and it begins to rain? Common sense tells a real human to take
shelter. However, in the case of the robot janitor, it doesn't know
that it's raining, unless you program it to take shelter when it
rains. Another example is what if the janitor accidentally drops
food on the ground; does it know that the food is contaminated?
This is why it is very important to build a machine that is similar
to a human brain in order for it to do anything human. The only
way to build such a machine is by making software that can understand
language.
[0173] Language is important because the robot needs to learn things
from a society. The only way that humans can communicate with robots
is if they both have some form of common language so that both parties
understand each other. People who speak English can understand each
other because the grammar and words used can be understood by everyone.
Think of language as the communication interface between human robots
and human beings.
[0174] There are basically 3 things that the AI software has to
represent in the language: objects, hidden objects, and time. I
don't use English grammar because English grammar is a learned thing.
These 3 things I mentioned are a better way to represent language.
If you think of objects as nouns and hidden objects as verbs, then
that is what I'm trying to represent.
[0175] Objects
[0176] One day when I was playing a game for playstation 2, I couldn't
notice that the game was repeating itself over and over again. When
the characters jumped the same images appeared on the screen. When
the enemies attacked the same images appeared on the screen. These
repeated images was what gave me the idea that I can treat all the
images on the screen like image layers in photoshop. I can use patterns
to find what sequences of images belong to what objects. When the
360 degree images of one object is formed then I can use a fixed
noun to represent that object (I call this 360 degree image sequence
a floater). For example, if I have the 360 degree floater for a
hat I can assign the letters "hat" to the floater. If
I have the 360 degree floater for a dog I can assign the letters
"dog" to the floater. The image processor will dissect
the image layers out and the AI program will determine what the
sequential image layers are. This is done by averaging the data
in memory--taking similar training data and analyzing what the medium
is. When the averaging is finished the floater has a range of how
"fuzzy" the object can be.
[0177] Things like cat, dog, hat, dave, computer, pencil, tv, book
are objects that have set and defined boundaries. Things like hand,
mall, united states, universe don't have set boundaries. Either
it doesn't have set boundaries or they are encapsulated objects.
One example is the foot, when does a foot begin and when does a
foot end? Since a foot is a part of a leg it is considered an encapsulated
object. Another example is mall, when does the mall end and when
does it begin? Since there are many stores and roads and trees that
represent the mall we can't say where the mall ends and begins.
The answer is the computer will figure all this out by averaging
the data in memory. Another thing is that some objects are so complex
that you have to use sentences to represent what it is. The Universe
is one example, when does the universe begin and end? The answer
is we use complex intelligence in order to represent the meaning
to the word "universe".
[0178] Unfortunately, black and white drawings are preferred in
utility patents so I decided not to use colored pictures of videogames.
(In U.S. Provisional Application No. 60/909,437, all examples are
demonstrated by videogames) Instead I decided to use black and white
images of animated movies and comic strips to illustrate my point
about objects, hidden objects and time.
[0179] The first two pictures in (FIG. 17 and FIG. 18) best illustrate
the point about image layers and floaters. The first picture 44
displays a series of lines and shapes that make up images. There
are many things that are displayed in the picture 44. There are:
the moon, the city, the tentacles, the walls, the characters, the
breakable objects and so forth. The image processor will dissect
the most important image layers from the picture (this process can
be done in black and white but the image processor will have an
easier time with colored pictures). It will then attempt to find
a copy of this image layer in memory. Based on certain patterns
within all the colored pixels and the relationship between each
other the AI will understand what image layers belong "sequentially"--consistency
and repetition is the key. The computer will normalize all the image
layers (including encapsulated image layers) until it comes to an
agreement of what is considered an object and what are encapsulated
objects. Referring to FIG. 17, in list 46 is an example of 3 major
image layers (objects) that the computer has found: Spiderman, Doc
oct, and the background.
[0180] The purpose of the image processor is not to identify the
image layers, but to delineate image layers that are moving from
one frame to the next. The identification of the image layers comes
by finding the image layers in memory. The image processor only
makes the search process much easier to identify the image layers.
One example is the Doc oct image layer. The image processor doesn't
know that the tentacles belong to Doc oct. In fact, the image processor
will think that the tentacles are separate image layers. Only when
the AI identify Doc oct in memory does the AI know that the tentacles
is a part of Doct oct.
[0181] Now that the image processor has found Spiderman 48 as one
image layer, it will randomly break up Spiderman 48 further into
partial data. This is represented by letters: M, N, O, P, Q, R.
The partial data will each be searched randomly in the network.
[0182] Although I couldn't find comic strips for Spiderman I found
comic strips for Charlie Brown instead. In FIG. 18 the image layers
of Charlie Brown are cut out from the movie animations 50 and 52.
On the second picture (FIG. 19) is the 360 degree floater of Charlie
Brown 54. All the possible moves of the character including scaling
and rotation are stored as sequences in this floater. If the movie
sequence is in 360 degree, like in a videogame, then the floater
will have 360 degree image layer for each possible outcome. If the
movie sequence is in 2-d then the floater will have only possible
outcomes of the character. "The creation of the floater is
kind of like reverse engineering a videogame programmers work or
reverse engineering an animators work--what do videogame programmers
consider an object or what are the animators' cell layers".
[0183] The next step is to take the floater and treat it as an
object. This is how I represent objects visually in my program--by
using patterns to find the 360 degree images of an object and all
its possible moves. The rules program 56 will bring the object "Charlie
Brown" and the floater of Charlie Brown 54 together (FIG. 19).
The target object is the word "Charlie Brown" and the
floater is the element object. Once the floater passes the assign
threshold that means the word "Charlie Brown" has the
same meaning as the floater. At this point, any sequence wither
its one frame or 300 frames of the floater is still considered the
same object. You can stare at a table for hours but the table will
still be a table. You can also walk around and stare at the table,
the sequential images you see is still a table. The question people
ask is: what happens if you break the table or what happens if there
are other objects that make up a table. The answer is the AI will
normalize the objects and output the most likely identification.
[0184] There are other topics that concern objects such as encapsulated
objects (a human object can have thousands of encapsulated objects)
and priority of objects and partially missing objects but I won't
get into those topics.
[0185] Hidden Objects
[0186] Sometimes there are objects that don't have any physical
characteristics. Action words are things that don't have physical
characteristics. Things like walking, talking, jumping, running,
throwing, go, towards, under, over, above, until, and so forth.
These words are considered hidden objects because there is no image,
sound, taste, or touch object that can represent them. The only
way to represent these objects is through hidden data that is set
up by the 5 senses. Let's call the 5 senses the current pathway--the
pathway that the computer is experiencing. In order to illustrate
this point I will only refer to the visual part of the current pathway.
[0187] Within the visual movie are hidden data that I have set
up. This is done because I wanted the computer to find patterns
within visual movies. Some of these hidden data are: the distance
between pixel/s and the relationship between one image layer and
another image layer. Let's illustrate this point by using a simple
word: jump. The computer will take several training examples from
the visual movie regarding jump sequences. As you already know,
variation to a jump sequence can range exponentially. A person can
jump from the front, back, side, at an angle, top, 10 feet away,
or 100 yards away. The person doing the jumping can be other objects
such as a dog, rat, horse, or even a box. There are literally infinite
ways that the jump sequence can be represented in our environment.
The computer will take all the similar training examples and average
the hidden data out. Every time that a hidden data is repeated the
computer makes that hidden data stronger (hidden data are considered
objects). The hidden data are also encapsulated so that groups of
common hidden data are combined into one object. As more and more
training are done the computer will have the same hidden data for
the same fixed word: jump. The rules program will bring the word
"jump" and the hidden data closer to one another. When
it passes the assign threshold the word "jump" will be
assigned the meaning (hidden data).
[0188] In FIG. 20 the picture 58 is an example of how the word
jump is assigned a meaning. First, the computer analyzes each jump
sequence: J1, J2 and J3. It will analyze all the hidden data that
all three jump sequences have and group those common traits into
an object. Then the rules program 60 will take the word "jump"
and assign it to the closest meaning.
[0189] The rules program is another thing I want to mention. When
you train the robot, timing of the training is crucial. The reason
why the word jump is associated with the jump sequence is because
the jump sequence happens and either during the jump sequence or
closely timed is the word "jump". The close training of
the word jump and the jump sequence is what brings the two together.
If the word "jump" is experienced and the jump sequence
happens 2 hours later, the computer will not know that there is
a relationship between the word "jump" and the jump sequence.
This is how the machine will learn language, by analyzing closely
timed objects. This is also a way to rule out coincidences and things
that happen only once or twice.
[0190] Time
[0191] Time is another subject matter that has to be represented
in terms of language. In my program there is no such thing as 1
second, 1 minute, 5 years, or 2 centuries. The time that we know
are learned time and isn't used in my program. What I have done
is create an internal timer that will run infinitely at intervals
of 1 millisecond. The AI will use this internal clock and try to
find if there are objects (words) that have relationships to the
internal clock. The timing in the AI clock can also be considered
an object. For example, if someone says "1 second". After
many training examples the computer will find a pattern between
"1 second" and 100 milliseconds in the AI's internal clock.
This internal clock of 100 milliseconds will be an object that has
the same meaning as "1 second".
[0192] The above information concludes how my program represents
things like nouns, verbs, time, and grammar. When we are dealing
with entire sentences the computer has to do all the hard work by
averaging all the training examples, looking for patterns, and assigning
meaning to words in the sentence. The sentence itself is considered
a fixed movie sequence while the meaning to the sentence changes
as the robot learns more In FIG. 16 the diagram gives an example
of how the rules program will assign meaning to the sentence "the
box jumped over the dog". Just like how the rules program learn
nouns and verbs, it will learn the meaning of the sentence by finding
the "complex patterns". The target object is broken up
into sub-groups and the element objects are broken up into sub-groups.
The AI will then attempt to string the element objects and combine
them into other element objects that best represent the entire sentence.
[0193] This type of machine to represent language is considered
"universal" because the program can be applied to all
languages including sign language. Different languages use different
words to represent the same things. "cat" in English,
"neko" in Japanese, and "mau" in Chinese are
all talking about the same object. Different verbs in English, German,
or Latin are all talking about the same verbs. Even something like
sign language uses fixed sequential hand motion to represent words
and phrases. The grammar too also relies on patterns and different
ways of stringing words/verbs together to mean something. This is
easily done with the AI program because finding patterns is what
it was designed to do. As long as the grammar in that language repeats
itself or have some kind of rule (regardless of how complex) then
the pattern will be recognized by the AI.
[0194] Patterns and Language
[0195] Now that I have discussed all the basics of how most words
are represented let's get into something more complex such as finding
patterns. When a question like: where is the bathroom? is asked,
patterns are used to answer the question. These patterns are found
by averaging similar pathways in memory. Some of the functions used
to find patterns include: using the 3-d environment (in storage),
using visual functions such as pixel comparison and image layer
comparison, using long-term memory, searching for specific data
in memory, and so forth. Where is the book, where is the sofa, where
is Mcdonalds, where is the University, where is dave? All these
questions rely on their respective universal question-answer pathway.
The AI will look into memory and find out that there is a relationship
between a question and a specific type of pattern to get an answer.
In terms of the bathroom question, the AI will find that it has
to know where it is located presently (this is done by looking around
and identifying its current location). Then the robot will look
into memory for the bathroom that is located in the current location.
If the bathroom location is found in memory it will output the answer:
"the bathroom is located -----". If it doesn't know (no
bathroom memory in current location) it will either say it doesn't
know or it will attempt to find more information to answer the question.
[0196] This pattern finding doesn't just apply to questions and
answers but also statements and orders. If someone said: "remember
to buy cheese at the supermarket". This statement has a recurring
pattern and it requires that there are many training examples so
that the AI can find these patterns. The pattern is when the robot
gets to the supermarket, sometime during the purchase of goods,
the statement pops up in memory "remember to buy cheese".
Sometimes the robot forgets (either a learned thing or the pattern
wasn't trained properly).
[0197] The data in memory will become stronger and stronger as
more training is presented. Language or sentences are considered
data in memory. These type of data will become considerably stronger
than other data because language is fixed while other things constantly
change. Language is what humans use to classify other data in our
environment which includes visual objects, nouns, verbs, sentences,
scenes, description, tasks, and the like. In other words, language
brings order to chaos. This is why when we take input from the environment
language has top priority over other data. This is also why our
conscious activates sentences and visual scenes more than anything
else when we consciously think.
[0198] The AI will average all the data in memory and create a
fuzzy range of itself called a floater. Data in memory would include
images, objects, pathways, entire scenes, and so forth. Averaging
of data (or self-organizing of data) takes place when input is stored
in memory. After the averaging, a fuzzy range of the data will be
the result. In terms of sentences the average meaning of the sentence
will be stored and not an exact sentence.
[0199] A. Averaging the Meaning of Sentences
[0200] When teachers say:
(Y1) "look left, right, and make sure there are no cars before
crossing the street"
(Y2) "remember to see if there are no cars from the left and
right before you cross the street"
(Y3) "don't forget to look at all corners to make sure there
are no cars before crossing the street"
[0201] All the sentences are saying the same thing. This is why
language is so important, we can interpret language infinite ways
and they are all talking about the same things. The computer will
recognize all of these things and it will average out what the meaning
of the sentence is
[0202] Referring to FIG. 25, after many training of the pathway
the AI has universalized the groups of pathways (Y1, Y2, Y3). Y1,
Y2, and Y3 disappear and what you have left is the average of all
the training data located in that area (Steps 86 and 88).
[0203] The AI not only averages out trees in pathways but entire
pathways. The purpose is to universalize similar pathways into one
pathway. This one pathway will contain the fuzziness of infinite
possibilities. We can also take this universalized pathway and encapsulate
that to make even more complex pathways.
[0204] The next two examples illustrate how language can be incorporated
into the human conscious to accomplish tasks and solve problems.
[0205] A. ABC block
[0206] B. Answering universal questions
[0207] ABC Block
[0208] In this problem we want to use a basic intelligent problem
that kids can solve. The ABC block is just 3 square blocks and the
robot has to find a way to stack the blocks in an A B C format.
[0209] We accomplish this problem with the English language. We
simple tell the machine: "I want you to stack the blocks up
starting with C then B and finally A". From this one sentence
the robot should be able to finish the task. It doesn't matter what
order the blocks are put in. It doesn't matter where the blocks
are. If the robot understands the sentence it will carry out the
command. Of course we have to train it to understand the steps to
accomplishing this easy task. Let's say that we had the blocks in
this order and we wanted the robot to stack the blocks up from ABC
(in FIG. 26)
[0210] Referring to FIG. 26, we learned from teachers that in order
to solve this problem we: "locate the C block", "Take
the C block and put it on the ground", "then find the
B block and put it on the C block", "finally find the
A block and put it on the B block". These sentences are trees
that tell you what to do in order to solve this problem. These trees
were trained by a teacher many many times before you can attempt
to solve this problem. By the way, these trees are your conscious
(Step 90, 92 and 94).
[0211] These trees encapsulate the instructions to accomplish a
goal. We train them by teaching the robot that this sentence is
followed by these instructions. The robot will create pathways in
memory that will store the instructions step by step. This may not
sound impressive but let's say you wanted to solve something like
lining up the entire alphabet letters in a certain order. If you
preprogram the solutions there will be couple trillion possibilities
you have to manually preprogram. With trees we can encapsulate instructions
in the form of sentences. And these sentences can be encapsulated
into even more complex problems, thus making a complex problem into
a simple problem.
[0212] Answering Universal Questions
[0213] The answering of questions relies on patterns in order to
be understood. We are able to find the patterns and universalize
the pathways so that when someone ask us a question we can give
them the appropriate answer.
[0214] 8=8 is an equal object or Dave=Dave is an equal object.
They are equal is the relationship between the two objects. Whenever
the computer finds two objects equal it will establish a relationship
between the two objects and find patterns that revolve around these
two objects. From (FIG. 27A-27C) we have taken all the equal objects
and we have tried to find patterns between those equal objects.
Answering questions is a pattern that relies on equality to find
the answers. This may not be very clear when you look at the first
example, but after looking at the second example and comparing that
with the first example there is clearly a pattern there.
[0215] By establishing a relationship between equal objects the
computer will be able to find patterns between different training
data and forge a universal pattern that can answer a universal question.
The examples in (FIG. 27A and FIG. 27B) have a pattern which is
depicted in (FIG. 27C). In FIG. 27A data 96 in memory is used to
establish equal objects to the sentence 98. In FIG. 27B data 100
in memory is used to establish equal objects to the sentence 102.
In FIG. 27C a pattern has been established represented by blocks
104 and 106.
[0216] The pattern found in (FIG. 27C) can answer any question
that has that kind of configuration. Examples of this would be:
what is 8+8? 8+8 is 16.
what is the 21st state in the USA? The 21st state in the USA is
Illinois.
what is the first letter in the alphabets? The first letter in
the alphabets is `A`
what is the last letter in the alphabets? The last letter in the
alphabets is `Z`
[0217] As you can notice that this whole human level artificial
intelligence program is all about finding patterns. I set up the
different kind of patterns to look for and the computer uses the
AI program to find those patterns and assign those patterns to language.
Language will always be fixed (unless society changes it) but the
patterns that represent language changes from one time period to
the next. There are also multiple meaning to fixed words.
[0218] The Relationship Between HLAI and the Human Brain
[0219] The data structure of a human brain and something like a
calculator are totally different. On one hand a calculator can process
thousands of equations each second but the human brain processes
only 1 equation per second. This doesn't mean that the calculator
is more superior than a human brain. It just means that the brain
is a different form of computer that processes information differently.
The human brain is a very powerful computer that can learn from
past experiences and understand common sense knowledge which is
something current computers can't do.
[0220] The human brain consists of 10 billion neurons and 60 trillion
connections. The data are stored in the neurons in terms of encapsulation
and commonality. Although the brain has only 10 billion neurons
it is able to store almost 8,000 trillion data because of the connections
that each neuron has with other neurons. The data are also global
in nature and each neuron will have associations with other neurons.
All of the neurons and their connections are either strengthened
or forgotten. The neurons get strengthened by a process of chemical
electricity that makes their connections with other neurons stronger
(or weaker).
[0221] When an object is recognized like an image or a sound, electricity
is run through that neuron and its connections (FIG. 21A). This
is how psychologists can understand what parts of the brain does
what functions--by using a computer to analyze the electrical activities
in the brain. Since there are many sensations coming into our brain
each second, there isn't just one area the brain is active but activity
will run in multiple areas of the brain at the same time.
[0222] I did some observation of how the brain sends electricity
throughout the neurons and came to the conclusion that we can actually
simulate this activity in a software. First the brain locates an
object (let's call this object the target object). In this case
an object could be anything--it can be an image of a car or a sound
of a dog barking. Once the brain locates the target object in memory
it runs electricity throughout all of the connections associated
with that object. This will strengthen not only the target object
that has been located but it will bring all the other objects (call
these element objects) closer to the target object.
[0223] When the AI locates the three visual objects: A, B, C in
memory it will run electricity through these nodes and all of its
connections (FIG. 21A).
[0224] Referring to FIG. 21B, the mind 72 has a fixed timeline.
Only one element object can be activated at a given time in this
timeline. This is how we prevent too much information from being
processed and allow the AI to focus on the things that it senses
from the 5 senses. Step 70 activates qualified element objects in
mind 72 in linear order.
[0225] This finding is important because we know that the target
object that the brain has located has to be strengthened. This is
done by applying chemical electricity through that located target
object. The only question I had was: "why did the electricity
propagate throughout all of its connections too?". Would that
not strengthen all the element objects around the target object
too?
[0226] The reason why the brain had to propagate electricity throughout
all of the target object's connections is because that is how the
conscious is presented. The conscious is the voice in your head
that speaks to you. It also gives you information about a situation,
or help you solve a problem, or tell you definition of words. Referring
to FIG. 21A, all the element objects 66 from all the target objects
64 will compete with one another to activate in the mind (the mind
can only take in a limited amount of information). When that information
is activated in the mind a lesser amount of electricity will be
applied to that information and its connections. This is how the
mind travels from one subject matter to the next.
[0227] The brain modifies information by constantly applying chemical
electricity throughout all the target objects coming in from the
5 senses (Step 68). The electricity strengthens not only that target
object but it strengthens all the element objects that have association
with the target object. This form of storing, retrieving, and modifying
information in a network is what allows the host to have human-level
intelligence. The next two paragraphs demonstrate how the conscious
works in terms of reasoning and interpreting grammar.
[0228] Reasoning happens when two or more objects recognized by
the AI share the same element objects. The more objects share an
element object the better the chance it will get activated. For
example, if you had a statement like:
[0229] If the weather is sunny and I have free time and my dog
is blue then go to the beach.
[0230] So, if the AI recognizes "the weather is sunny"
and "I have free time" and "my dog is blue"
then the stereotype will activate: "then go to the beach".
The recognizing of the objects can also be in any order. These objects
can also be a fuzzy range of itself such as the statement: "I
have free time" can be represented as "I don't have to
work today".
[0231] Understanding entire sentences, which was discussed earlier,
depend greatly on the conscious. Understanding grammar structure
of a language will depend on things learned in the past (FIG. 5).
For example, how are we supposed to learn a word like: jumped. The
word jumped has an ed at the end and we know from English classes
that if a word has ed at the end that means the verb (jump) happened
already. So, when the AI encounters a word like jumped the conscious
tells the AI that "words with ed at the end means the jump
happened". This is an element object that activated when encountering
the word jumped. This element object tells the AI what the meaning
of jumped is.
[0232] Predicting the Future
[0233] The main function of the HLAI is to predict the future based
on the current event. When the AI is applied to a car the current
driving state is the current event. The AI has to predict the future
so that it can steer the car in the right direction. Out of all
the pathways in memory the machine can only follow one given pathway,
the optimal pathway. This optimal pathway represents the best pathway
the AI can follow to act intelligently in the future. Predicting
the future isn't a very easy thing to do. In order to do that the
AI must first determine the worth of each pathway in memory based
on two criterias: the closest pathway matches and calculating the
worth of their future pathways.
[0234] The next couple of paragraphs are a recap of how the AI
program predicts the future. In (FIG. 1) the program has one for-loop
that repeats itself over and over again. The idea is: The computer
takes in one frame from the camera, it calculates the best possible
future to take, then it takes action. The computer takes in one
frame from the camera, it calculates the best possible future to
take, then it takes action. The computer takes in one frame from
the camera, it calculates the best possible future to take, then
it takes action. This loop repeats itself over and over again until
the AI is shut down (the instructions in the for-loop must be accomplished
within a predefined time limit, usually 1 millisecond). Human beings
work pretty much the same way, we take in input from the environment,
the brain calculates the best future course, then the human being
takes action. This repeats itself over and over again.
[0235] In FIG. 1, the first step is to search the current pathway
in memory for the closest matches (Step 4). The computer will list
the ranks of the searches starting with the best match (Step 6).
Next the AI will find future pathways for each of the matches and
calculate their future prediction worth. Then, the AI will decide
based on the matches and the future prediction on which pathway
is worth the most (Step 8). Finally, the AI chooses one pathway
to follow (Step 10). This one pathway is the optimal pathway and
it will be used to control the AI.
[0236] In FIG. 2, I show how the function works from a different
angle. The computer basically matches the current pathway with the
best match in memory then it calculates the best possible future
to take.
[0237] This form of artificial intelligence method to predict the
future has not been explored before because the possible outcome
of an event in life is infinite and the computer can't store all
the possibilities in memory. In order to drive a car the AI has
to store all the possibilities of driving a car in memory. This
would be impossible because the variations of life are infinite
(can you imagine storing infinite hours of driving in memory?).
This is why researchers have abandoned this field of AI. In my program
I made it so that the movie sequences are stored in a fuzzy logic
way. The most important data are kept and the least important data
are forgotten. This will allow the AI to anticipate the most likely
outcome of an event. Self-organization knits all the data together
forming object floaters in memory so that one given data has a fuzzy
range of itself. One example is a cat. A cat can come in all different
kinds of shape, sizes, and color. The strongest sequential images
of a cat are considered the center of the object (floater). After
determining a predefined range of how fuzzy the cat object (floater)
can be, anything that falls within this fuzzy range will still be
considered a cat object. The AI will be able to take in any picture
of a cat, regardless of how distorted or different it may be, and
still identify it as a cat. This is how my program can store infinite
amounts of data, by taking the average of an object and creating
a fuzzy range for that object. Object floaters don't just apply
to individual objects like cat, dog, or shoe, but entire situations
or language. Every data in memory has a fuzzy range of itself. The
next several paragraphs demonstrate how fuzzy logic is used to predict
the future for similar or non-existing pathways in memory.
[0238] When my computer program doesn't find a 100 percent match
in memory the AI has encountered a deviation (finding a 100 percent
match is very rare). There are 4 deviation functions I have set
up to solve this problem. It will allow the future prediction to
do its job properly and find the most likely next step. I will be
using videogames to illustrate this point. Videogame colored pictures
can't be used so the images will be done with animated movies. The
4 deviation functions are:
[0239] A. Fabricate the future pathway based on minus layers.
[0240] B. Fabricate the future pathway based on similar layers.
[0241] C. Fabricate the future pathway based on sections in memory.
[0242] D. Fabricate the future pathway based on trial and error.
[0243] Fabricate the future pathway based on minus layers
[0244] In FIG. 22 the AI minuses layers from the pathways and finds
the commonalities between the current pathway 50 and the pathways
in memory. For videogames/animation the AI minus object layers from
the game. The background layer is minused from the game and the
remaining layers matches the current pathway 50. This means the
sofa, the blanket, the walls, snoopy, and the captions are minused.
The two character layers (Charlie Brown and his friend) are used
to play the game (pathway 74).
[0245] Fabricate the Future Pathway Based on Similar Layers
[0246] In FIG. 23 the AI will find similar layers between the current
pathway and pathways in memory. For videogames/animation the AI
finds similar object layers. The Charlie brown layer with the hat
(pathway 76) isn't stored in memory. However there is a similar
Charlie brown layer without the hat stored in memory. Because the
Charlie Brown layer with the hat (Pathway 76) and the Charlie Brown
layer without the hat (Pathway 78) look similar the computer will
use Pathway 78 instead of Pathway 76 to play the game.
[0247] Fabricate the Future Pathway Based on Sections in Memory
[0248] In FIG. 24 the AI constructs new pathways from sections
in memory. This process takes sections of pathways from memory and
combines them to form new pathways for the AI to pick. Pathway1
is the pathway it is looking for in memory. However, there is no
100 percent match in memory. The closest match is pathway2. It takes
section1 and section3 from pathway2 and fabricate pathway3. This
fabricated pathway will be used to play the game.
[0249] Fabricate the Future Pathway Based on Trial and Error
[0250] The AI plots the strongest future state and fabricates a
pathway to get to that future state using the other deviation functions.
[0251] With all 4 deviation function the AI program can fabricate
pathways in memory if there are no exact matches found. All four
deviation functions create the fuzzy logic of the system. It acts
by giving the AI alternative pathways if an exact match isn't found
in memory. It also gives the AI the ability to predict the future
of pathways that are similar or non-existing in memory.
[0252] For future predictions, the weights of future sequences
in the pathway has already been established by training and only
require the AI to predict 3-4 steps into the future to receive an
accurate prediction of thousands of steps into the future. In some
cases future prediction isn't required because of this system to
store/retrieve and modify information (FIG. 14).
[0253] The steps to calculating the worth of future pathways are:
designating a current state in a given pathway and determining all
the future sequences in the pathway; adding all the weights for
each possible future sequences; calculating the total worth of each
possible future pathway and ranking them starting with the strongest
long-term future pathway (search algorithms such as A*, hill-climbing,
depth-first search, breadth-first search, iterative deepening A*
can be used to search for future pathways).
[0254] Long Term Memory
[0255] One other subject matter I will discuss is long-term memory.
Long-term memory is just one long computer log of sequential movie
events collected by the AI. The long-term memory is actually a timeline
with references to sequential data collected by the AI (in increments
of 1 millisecond). When the data in the network is forgotten the
data in long-term memory is also forgotten. However, the forget
rate isn't as smooth and linear as a straight line. The remembering
of data is based on emotional factors, pain or pleasure, the AI's
intelligence level, and other innate factors such as attractiveness
or ugliness. Memory will be forgotten centered at the current state;
the farther the data is from the current state the more it forgets.
This doesn't mean that data 10 years ago is less clear than data
1 week ago. Sometimes data that happened 10 years ago is stronger
than data that happened 1 week ago because the AI has a strong recollection
of an event or that data is being recalled many times by the AI.
[0256] Finding patterns is the single most important trait used
to produce human level artificial intelligence. The long-term memory
is used in the pattern finding process. The 3-d storage and the
3-d environment are also used in the pattern finding process; along
with thousands of other embedded data or functions. This part of
the program is very complex and long and is beyond the scope of
this present invention. The most important patterns are disclosed
in this patent.
[0257] The long-term memory has embedded data in it to help the
AI find patterns. Having the ability to rewind and fast forward
movie sequences to find information is a valuable asset. For example,
if someone wanted to know when the AI machine saw a car accident,
the machine will use the long-term memory to locate the time it
saw the car accident. If someone wanted to know how long it took
the machine to finish a task the machine will locate the movie sequence
that contain the task and give an approximate time it took to finish
the task.
[0258] The 3-d storage which maps out a 3-d environment has embedded
data in it to help the AI find patterns. For example, if someone
wanted to know where the closest Mcdonalds is in a city, the machine
has to look in the 3-d environment (3-d storage) and locate where
the city is and the closest Mcdonalds is. If someone wanted to know
the approximate distance from one location is to another location,
the machine will use the 3-d environment to find the approximate
distance.
[0259] All these patterns are found on its own through observation
and learning. No fixed rules or policies are needed to learn how
to do things. Answering questions is learned on its own, finding
out solutions to problems are learned on its own, learning the rules
of driving a car is learned on its own, and so forth. There are
no predefined rules to tell the AI what to do and what not to do,
everything is learned from society.
[0260] Learning from Childhood to Adulthood and how the Pathways
Become More Complex
[0261] When the machine is at its early stages of life, it will
have to build its pathways from simple data then as it gets older
and there are more data in memory it will organize the pathways
into complex intelligence. Just like how we humans have to learn
to walk, to talk, to move, to eat, these machines have to go through
life the same way. Let's illustrate the gradual forming of simple
data into intelligent data by outlining a series of stages.
[0262] 1. innate reflexes
[0263] 2. trained to do things
[0264] 3. sequential events
[0265] 4. sentence commands
[0266] 5. give robot option commands
[0267] 6. practice makes perfect
[0268] 7. copy other peoples behavior
[0269] 1. Innate Reflexes
[0270] In this stage the robot will learn all the different objects
that are in the environment from the 5 senses. Things like cat,
dog, table, chair, red, blue, car, house, I, her, him, loud, soft
etc. are learned and stored in memory. The 3-dimensional floater
of all the objects will be created. Then the robot will start to
move its arms and legs from innate built in reflexes. Movement of
the arms, the legs, movement of the mouth, and controlling the vocal
cords are the things that the robot must learn first. These experiences
must be stored in memory in an organized way. Curiosity will be
the factor that steers the robot into doing things that it never
did before. Things like new objects it never learned before will
have top priority over old objects it learned. New sensations will
be more focused on then old sensations. By the time the robot learns
most of the objects around him its memory banks will be filled with
data and things around the robot will be more familiar. Meaning
of the objects will also be established.
[0271] 2. Trained to do Things
[0272] This part is where a teacher will guide the robot into doing
things that are appropriate and to force the robot to learn things
that it supposed to know (FIG. 28A). Things like walking, and grabbing
object, and throwing things around must be learned. The guide is
used so that the robot will learn important things that it can use
to control the environment. A thing like walking is important because
we want to get from one destination to another. Writing using a
pencil is important because we must learn to write letters. Things
like walking and writing and speaking must be learned by a guide
because we can't preprogram the robot to learn these things.
[0273] Although the guide isn't something we want to store in memory,
the point is that the more we guide it the stronger the desired
created pathway will be (referring to FIG. 28A). When it is strong
enough it can be used by itself and the guide pathway will be forgotten.
The robot will find a way to use the desired created pathway to
accomplish a goal. Walking for example, if the robot knows that
walking will get it from one destination to the next, then when
it sees food, it will use the walking path to go from its current
location to the food. Reward is also playing a part in this learning
process.
[0274] Also, during this process simple sequential consequences
will be understood. Things like what is the consequence of dropping
a ball, where should the ball be when you drop it, and solid objects
and soft objects have different properties.
[0275] 3. Sequential Events
[0276] In this stage the robot begins to learn how objects interact
with one another. When two objects hit each other both objects suffer,
when the robot fall down it's painful, when it grabs a solid object
it has the same shape, but if it grabs a soft object it bends its
shape. So, sequential events will be learned. The consequences of
the robots actions in comparison to the environment will also be
learned. By learning all these things the individual data in memory
will turn more complex and long. The robot will be able to piece
together the outcome of an event just by looking at its past. Another
thing to remember is that curiosity is the key to new pathways.
The more unique the event is the more the robot wants to learn it.
The old events it learned many times will be ignored because it
learned it already, but the new sensations will guide it to learn
new things. Think of curiosity as a form of pleasure and old sensation
as pain. Since this robot does things in terms of pleasure it will
look for new data from the environment. At this stage things like
lying and magic can't be distinguished yet. The robot will not be
able to lie yet and if it sees a man flying in the sky or walking
on water the robot will think it is real.
[0277] 4. Sentence Commands
[0278] This part will require the robot to know basic grammar like
the names of most objects that are around the environment. These
basic grammar must be thought to the robot and understood by the
robot. The rules program will do the rest by assigning the meaning
for the grammar. Even hidden objects must be understood like jump,
run, walk, loud, soft, etc. Once a basic language is established
we can combine sequential events with grammar and force the robot
to do things by using words as the tool. An example would be if
you said sit and the robot sits. When you say: "pick up the
book" the robot will pick up the book. When you say: "read
the first paragraph" and the robot reads the first paragraph
of the book. These are commands that you give to the robot to indicate
what you want it to do. There is no deception, or lying involved
in the command process. It's simply someone giving a command and
the robot taking the action. The robot may not understand what you
said and make a mistake, but having a voice in the head that tells
the robot to do things hasn't been created yet.
[0279] 5. Giving the Robot Option Commands
[0280] This part is an extension of the last stage. Instead of
saying a word and letting the robot do things we can add trees to
the command pathways and let the robot decide what it wants to do
(FIG. 28B). This is very affective because trees combined with commands
allow the robot to use if statements to accomplish a goal.
[0281] So, the tree decides what the robot will do. If a teacher
gives the command then the robot will listen, if it's a friend that
gives the command the robot won't listen. There are also innate
likes and dislikes the robot will have and there are commands out
there that tap into that kind of thing. For example, if the robot
was given this command: "pick the food you like to eat".
Within the robots memory there are powerpoints that determine an
objects worth. PM will tap into that and pick the one with the most
powerpoints. Commands like: "pick the color you like",
"eat the food you like", "play with the toy you like",
"buy the present you want", "wear the clothes you
love", and so forth will all depend on the robot. These likes
and dislikes can also be a learned thing.
[0282] 6. Practice Makes Perfect
[0283] Now, let's get on with a more complex way the pathways can
be formed. When we practice something like riding a bike, we are
actually creating new pathways to ride the bike. Practicing will
help the robot to decide the best newly created pathway to pick
to accomplish a goal. We can build a pathway in memory that will
treat practicing something as a command.
[0284] Referring to FIG. 28C, this example shows that by using
English we can guide the robot to do infinite amounts of tasks.
This example is a practice pathway. It uses a command that will
tell the robot to do something until a desired outcome is present.
If it doesn't accomplish the goal then it will repeat itself until
the task is completed. At the same time this is happening more trees
can be added to this practice pathway, like, if you practiced for
7 times and you still didn't accomplish a goal then quite. Or when
you are hungry and you don't have the strength to shoot then stop
practicing. The existing pathways will add, strengthen, or minus
trees from it as the robot learns more. Instead of following commands
there are other factors to consider before you take action to accomplish
the commands. The robot will do the things that a society will consider
appropriate at the time. If a society says it should lie in order
to not do the task then that's what the robot will do. If a society
says the command isn't appropriate in this type of situation then
the robot will not follow it. If the robot finds the command dangerous
and it can really damage itself, then it will not carry out the
command. This is where the inner voice that is the core of the consciousness
is built. The consciousness is the average of the things thought
to the robot by society.
[0285] 7. Copy Other Peoples' Behavior
[0286] This part is a very powerful tool used to learn things.
We can go ahead and train a tree that will allow the robot to copy
certain things from what it sees (FIG. 28D). Things that it sees
on TV will be learned and copied by the robot. Copying will allow
the robot to learn the most appropriate things to do in a society.
When it is in a situation it will do things in terms of what society
as a whole did. The way it dresses, the way it behaves in school,
the things that it likes/dislikes, how to dress, how to take care
of itself, how to get money, how to get food to survive, what to
say to certain people, how to make friends, how to get good grades
in school, and finding answers to questions. All these things are
pathways that were learned by copying other people in our environment.
[0287] This part will require not only trees but also relations
to past data and innate instructions of the robot. Pattern matching
will find these hidden things and put them in the pathways. Something
as complex as copying people require that you understand the relationship
between the robot and other objects. If other people move their
hand, you will copy them by moving your hand. You would need to
know that your hand is one object and it belongs to you as an individual
and that the other person you try to copy has a hand too and they
are an individual too. Also, you have to understand when to copy
them. If a copy is one second after you see the person do the action,
then one second is the time it takes to copy their action.
[0288] From all these pathways we can build on each other and make
even more complex thinking such as representing a hierarchy system.
Things like parent-child relationships, who is the grandfather of
the family, or what does having a brother really mean, will be represented
by complex thinking. When people say "that's your father",
there are lots of complex things we need to know before we can understand
that kind of thing. Complex things such as: "where do humans
come from?", or "parents are supposed to take care of
their kids" or "everyone has one female parent and a male
parent" or "the male parent is the father and the female
parent is the mother". It is a very complicated intelligent
system when it comes to representing a family tree and in order
to understand it we must first learn the simple things.
[0289] Training Pathways
[0290] The AI program records all the sequential movie frames in
a timeline called long-term memory. Long-term memory also has reference
points to all data (sequential frames and its encapsulated format)
stored in memory. The sequential frames and its encapsulated format
are broken up into sections and stored in different parts of memory
depending on "what optimal pathways the AI program decides
to pick".
[0291] FIGS. 29-33 are diagrams to demonstrate how the AI program
creates templates and how the templates are trained in memory. The
training data for each iteration of the for-loop is known as a "template"
(FIG. 29). Each template has its own encapsulated format (FIG. 30).
[0292] The templates are used to train data in memory in a streaming
continuous manner where the AI jumps from one section of memory
to the next to identify, store and modify information in memory.
[0293] The whole process of storing data in memory and remembering
long information comes from a simple concept. We have to build a
storage area that would lengthen the pathways as it learns more.
This can be accomplished by templates.
[0294] The process goes like this: first we have to create templates
for the pathway we want to store in memory, current pathway (FIG.
31). Then we use the AI program to find the most optimal pathway.
Referring to FIG. 32 and FIG. 33, remember optimal pathways have
3 different types of pathways: sequential pathways, minus layer
pathways, and fabricated pathways. According to the follow pathway
(the pathways the computer decides to take), we store the templates
in those areas (Block 108).
[0295] The Template Residue
[0296] FIG. 34 are diagrams to demonstrate how templates are used
to lengthen pathways in memory. The way the pathways remember long
sequences is by the template residue. When the AI program jumps
from one pathway to the next it leaves behind template residues
in both pathways--the pathway it jumped from and the pathway that
it jumped to. These template residue lengthens a pathway.
[0297] For example, let's take an easy example like Section1 and
Section2 from FIG. 34. If the AI program decides to jump from Section1
to Section2, then Section1 should have some template residue 112
of Section2 and on the other hand, Section2 should have some template
residue 114 from Section1.
[0298] The more template residue section1 storage area has of section2
then the longer section1's pathway is. When the training reaches
a certain point section1's storage area will have a sequential pathway
to section2 in its storage area. In other words, the length of section1
has increased to include section2 in its storage area. This is how
the length of pathways get longer and longer.
[0299] The idea behind template residue and lengthening pathways
is to prevent the AI from jumping from one section of memory to
the next to find information. Also, to knit the entire data in memory
so that most likely sequences are stored in the same area. This
will prevent repeated pathways from being stored in memory. If two
sections in memory have a copy of where it came from then one of
the two pathways will eventually have a copy of both locations.
The dominant pathway (with the strongest powerpoints) will have
a permanent storage area of both pathways while the weaker pathway
will forget. The next time the AI encounters the same situation
or similar situation it will travel on the dominant pathway and
will not jump to other sections in memory.
[0300] FIG. 35A-35D are examples to demonstrate how templates are
used to lengthen pathways in memory. Notice that after encountering
the same situation 3 times section1 has both the pathways that were
originally separated in different parts in memory. Section1 remains
in memory because that is the dominant location for that sequence,
while section2 will eventually forget and only parts of the pathway
remain (FIG. 35D). When the AI encounters this situation for the
fifth time the AI will pick section1 as the optimal pathway to follow
(it won't jump around in memory from pathway to pathway).
[0301] Retraining Objects or Templates
[0302] As I mentioned earlier, templates, pathways, and floaters
are just objects. When we retrain the templates (example from FIG.
35A-35D), we aren't just training all the templates, but we retrain
the templates and its encapsulated format in terms of priority.
During the training phase the computer has only a certain amount
of time to retrain the data before times up and the training stops.
The important thing is that we should train the objects with priority
first then train those that have less priority.
[0303] The priority of the object is discussed in later sections,
but the point is that from all the data in the current pathway we
break up the objects into priorities. Then we find each master node
of the object and then we train the storage area with the object's
templates.
[0304] FIG. 36 is a flow diagram depicting the process of how objects
are trained in memory. There are millions and millions of same objects
in memory. Remember that I said that all data in memory is global.
Well, when an object is identified it must locate its master node.
When that master node is located, it will be retrained and this
master node will retrain all the sub-nodes that depend on the master
node. Because the master node was retrained all of its sub-nodes
are also retrained. This is how data in memory is considered global
and not individual. One same object in memory has profound affects
on other same object in memory.
[0305] How to Get Meaning and Stereotypes from Objects
[0306] FIG. 37 is a diagram depicting the structure of repeated
objects in memory. As you have no doubt noticed all same information
is interconnected and anything that has association to the information
is interconnected. The reason is because all data has a master node
116. This master node 116 has connections to the sub-nodes throughout
memory. If one sub-node is changed a signal will be transmitted
to the master node 116 and it will be changed. When the master node
116 is changed all the sub-nodes are changed too because each sub-node
have a pointer 118 to the master node 116. This system is very important
because now we can get meaning/stereotypes (element objects) from
not only the strongest node (master node 116) but the rest of the
sub-nodes too.
[0307] Referring to FIG. 37, in the case when a sub-node 120 is
requesting for stereotypes, it will first identify the master node
116 and the master node 116 will determine which pointers are strong
and which are weak. Usually the most recent created pointer connection
is the strongest connection and it contains the strongest meaning/stereotype.
All these different same nodes throughout memory will compete for
their respective meaning/stereotypes to activate. How much of the
stereotypes will be activated will depend on how long the robot
was focusing on the object. This competition will also be fought
with other object nodes and their stereotypes.
[0308] Advance Version of the Rules Program
[0309] FIG. 38 are diagrams depicting the rules program. The rules
program is designed to bring association between two objects in
memory. The more association two objects have the closer they will
be from each other (their connection weights become stronger). If
two objects are close enough they are considered equal and both
are declared the same object. The assign threshold is a radius centered
at the target object to indicate that any element object that passes
the assign threshold is considered equal to the target object. Other
element objects that fall outside of the assign threshold and have
association to the target object are either stereotypes or trees.
[0310] The human conscious works by identifying target objects
from the current pathway and using the rules program to activate
closest element objects from the target object. The key here is
that there are many same target objects in memory (FIG. 39A). The
rules program has to track the strongest copies of the target object
from memory. Then the rules program will take the element objects
from all the copies of the target object in memory and decide which
of the element objects to activate (FIG. 39B). The strongest copy
of the target object is the master node.
[0311] From all the same target object copies in memory the AI
has to extract their respective element objects and all the element
objects will compete with one another to be activated. The element
object with the strongest association will be activated (FIG. 39B).
[0312] This means that the AI program finds the meaning to a word/sentence/or
object in a global fashion. The entire network must be searched
in order to find the meaning to an object. This technique not only
works for the meaning of words/sentences/or objects but the stereotypes
of the word/sentence/or object. The self-organization is there to
bring common objects together so that repeated data is brought to
a minimal.
[0313] Details on what is Being Trained in Memory
[0314] The current 5 sense pathway (FIG. 40A) will store not only
the 5 senses that are coming into the AI, but the conscious thoughts
that are activated by the AI. Both types of data are crucial for
many functions including recalling information and finding patterns.
[0315] FIG. 40B demonstrates that the current 5 sense pathway stores
the 5 senses along with the activated conscious thoughts. The visual
representation of A B C are the 5 senses (visual) and the sounds:
"horse", "sun", and "tree" are the
learned groups. As the AI recognizes and identifies `A` from memory
the sound "horse" gets activated. When the AI recognizes
and identifies `B` from memory the sound "sun" gets activated.
And when the AI recognizes and identifies `C` from memory the sound
"tree" gets activated.
[0316] In FIG. 40B, objects above the timeline are from the 5 senses
(target objects) and the objects on the bottom of the timeline are
activated element objects. Visual `A` and the sound "horse"
are equal because they both are stored in the same assign threshold
(very strong association). This means that the letter `A` and the
sound "horse" are both one and the same objects. On the
other hand, stereotypes and trees that get activated are related
to visual images ABC, but are not the same objects. "that is
jon's horse", "that hurt my eye" and "look away
from the sun" are either trees or stereotypes activated based
on the visual images ABC.
[0317] This is very important to how the AI stores information
in terms of "fuzzy logic" instead of storing information
exactly as the AI interpret the information. Because such information
is so complex I'm going to show some simple examples to give the
reader an idea why I had to store information in this manner.
[0318] FIGS. 41A-41D are different examples of the ABC block and
how to solve this problem in terms of "fuzzy logic". I
have given three examples of the same problem but different situations
and different sentences (FIG. 41A-41C). Visually, the same problem
will look very differently--this problem can be in a classroom environment,
it can be watched on tv, or the setting can in a stadium. The one
thing that binds all these examples together is language. Like I
said before language brings order to chaos and is very important
to the development of complex intelligence.
[0319] All three examples of the ABC block problem are very similar
(FIG. 41A-41C). In fact, the instructions to accomplish the task
are identical. The only difference is that people use different
sentences to mean the same things. As discussed in previous lessons,
the meaning to language are considered hidden objects. The AI uses
patterns to find the complex meaning to language and assign a hidden
object to the sentence. Hidden objects are also encapsulated and
therefore subject to forget. Within all the complex patterns in
the encapsulated hidden object are common traits shared by same
sentences. These common traits are grouped together and it defines
what the language means in a fuzzy logic way.
[0320] In FIG. 42, letters A B T are the common traits (meaning3)
for both meaning1 and meaning2, so they will be grouped together
as one common trait. As self-organization occurs in the storage
area, common traits will be pulled closer to one another. The common
traits will be grouped together within multiple encapsulated hidden
objects in meaning1 and meaning2. As the AI learns more and more
these common groups get stronger and stronger. This will then create
a universal hidden object represented by meaning3. That meaning3
can be represented by infinite sentences that will mean the same
thing.
[0321] As the AI learns the same scenes over and over again, the
sentences used in each learning scene are different but the meaning
to the sentence remains the same. This will allow the AI to average
the sentence that is used in each situation (sentences used in real
life are different everytime). The only thing that remains is the
meaning of the sentence. Because the meaning and the sentence is
one and the same object, even though the exact sentence disappears
from memory the meaning remains (thus the sentence is not actually
deleted from memory).
[0322] The patterned sentence is actually the average of all the
similar sentences. The computer found a universal pattern to the
sentence that correlates with the meaning of the sentence. This
will allow the AI to understand infinite possible variations of
the sentence. For example, the sentence: "put R1 on the ground".
R1 is a variable that can be anything.
[0323] As a result of self-organization all three examples (FIG.
41A-41C) have been averaged out and a universal pathway is created
(FIG. 41D). This universal pathway to solve the ABC block can now
be used to solve this problem under "any" circumstances.
It doesn't matter where the blocks are, it doesn't matter what the
blocks look like, it doesn't matter where this problem takes place.
The problem can be solved under any circumstances.
[0324] Although an exact pathway match would be preferred instead
of the universal pathway, life doesn't work that way. Life is dynamic
and humans don't sense and interpret things exactly the same way
twice.
[0325] Another consideration is timing of the problem. The three
examples in FIG. 43 can be different lengths. One can be 10 minutes,
another can be 7 minutes, and the last one can be 15 minutes. The
timing will also be averaged out and there is an approximate time
that certain tasks has to be accomplished (the average timing of
certain accomplishment of tasks is also used to find complex patterns
to intelligence).
[0326] The final topic of this section is the decision part of
the AI program related to this ABC block. FIG. 44 is a diagram showing
decision making by the AI program. The AI was designed to find the
best match in memory. However, just because there are higher pathway
matches in memory the AI will not always pick the highest percent
match. The powerpoints of the pathways are also a big factor when
considering which pathway to choose. For example, if the universal
pathway for the ABC block is considered a 20 percent match to the
current pathway with a very high powerpoints and there is another
pathway that is 85 percent match but has a very low powerpoints,
then the AI will pick the 20 percent match instead of the higher
percent match because the powerpoints overshadow what is actually
being sensed.
[0327] This type of decision making makes sense if you think in
terms of the human conscious and not what you actually sense from
the environment. In very complex intelligence the majority of decision
making isn't based on the 5 senses. Decisions are based on what
you have learned in the past.
[0328] Self-Organization Using Both Learned Groups and Commonality
Groups
[0329] Both the learned groups and commonality groups must co-exist
in the same storage area. This means that commonality groups that
have 5 sense traits are grouped in the same general area, but at
the same time groups that are learned to be the same but are totally
different in terms of 5 sense traits are also grouped together in
the same general area.
[0330] One example of this is the face. The face is a learned object
because it's a word that represents a group of visual images. The
face encapsulates other learned objects such as words like: eyes,
nose, mouth, ears, hair, chin, cheeks, and eye brows. For each of
these learned objects are their respective infinite variations in
terms of visual images.
[0331] The learned groups guide the commonality groups to be stored
in one area. For example, if you have real-life face images of two
humans--a female and a male, and you have a face image of a cartoon
character (such as Yugioh), these images are totally different from
each other in terms of physical appearance and measurements of things
like eyes, nose, mouth, hair color, and so forth. However, the fact
that all three images is a face is what groups them together. The
learned group "face" brings the three images closer to
one another. Within this learned group the commonality group will
also self-organize and bring images with common traits closer together.
In the case of the three face images, the female human face and
the male human face will be closer together, while the cartoon face
is farther away.
[0332] FIGS. 45A-45D are illustrations showing how learned groups
and commonality groups organizes face images. On the first example
(FIG. 45A) the picture is an anime character 122. Notice that an
anime character 122 has eyes larger than a human, the nose takes
the shape of a triangle and the mouth is a small line. These visual
images do not correlate with the face of a human being. However,
because we learned that these visual images are classified as a
certain word (eyes, mouth, hair, nose, face, etc.), we group them
as the same (learned groups 120).
[0333] On the second example is a face of Yugioh (FIG. 45B), the
popular kids cartoon. Just like the first example (FIG. 45A) all
the major parts of the face is classified in terms of learned words.
Although the eyes deviate from what we would call eyes on a human
we learned that that image is an eye. The first two examples (FIG.
45A-45B) have very similar visual traits: the eyes are large, the
nose is a triangular shape, and the mouth is a horizontal line.
[0334] In example 3 (FIG. 45C) the same technique is being applied.
The robot face looks different from a human face. But, because we
identify certain images belonging to certain English words then
that particular image belongs in that word group.
[0335] As the AI learns more and encounter more and more faces
it will have an easier time classifying that image in which groups.
From the three examples in FIG. 45A-45C the first two faces (anime
and Yugioh) will be grouped together closely, but the third face
(robot) will be farther away (FIG. 45D). This is how the storage
preserve both learned groups and commonality groups together in
the network. This would also help tremendously in terms of searching
for information in the network because all the data are organized
in an encapsulated fashion.
[0336] All these learned groups (encapsulated or non-encapsulated)
do not have to be activated by the rules program. Sometimes the
conscious activate something else that is considered a learned group.
It activates a learned group without even thinking. In FIG. 45D,
all the AI needs is the learned group "face" to activate
and every image in the face falls into learned groups that are contained
in the "face" group. The image of the eye will be in the
"eye" group without being activated, the image of a nose
will be in the "nose" group and so forth. The "face"
learned group was there just to identify an approximate location
in memory. The self-organization does the rest of the work. These
things are done at an unconscious level. The one sound "face"
or an identification of a face (hidden object--learned group) is
all that is needed to store the image of a face and all the encapsulated
images in the face image in its respective learned groups.
[0337] Averaging Data (Floaters) in Memory
[0338] The AI program will learn things from its environment and
store all the data according to the configuration of data in memory.
The 3-d environment is created because the things we see around
us stay the same all the time. Most of the images we see stay the
same. This is important because memory forgets the temporary objects
and remember the permanent objects--things that stay the same all
the time. The 3-d environment will be created in memory because
the environment (majority) is fixed.
[0339] What about objects that don't have a permanent fixture in
memory and moves a lot? The answer to that question is that the
computer tries to self-organize all copies of that object in memory
and give the object an average location in memory. When we see moving
cars, people walking, and shows on television, we are actually storing
those sequences in that particular 3-d environment. FIGS. 46A-46F
are illustrations demonstrating how moving objects self-organizes
in memory. Referring to FIG. 46A, if we are at the supermarket and
we see George Bush 132, we are actually storing the movie sequence
of George Bush in the supermarket area in memory. Next, if we go
to the beach and we see George Bush 130, we are actually storing
the movie sequence of George Bush in the beach area in memory. Finally,
if we go to the library and we see George Bush 128 we are actually
storing the movie sequence of George Bush in the library area in
memory. This gives us 3 areas in memory that we have encountered
the object: George Bush.
[0340] In FIG. 46A, B2 represents the area the AI encountered George
Bush. Notice how close B2 is between the library and the supermarket.
Self-organization will knit B2 together and average out the storage
area. The B2 on the Beach is too far and self-organization can't
bring that part of B2 closer to the other two copies of B2. After
many training of data in memory B2 will have a more permanent location.
[0341] In FIG. 46B are two copies of B2 in memory. The B2 from
the library and B2 from the supermarket are close so they merged
into one object and both of the powerpoints from both copies are
combined.
[0342] In a more dynamic environment and there are many moving
objects the computer does all the hard work to self-organize data
and determine where to store the object. Objects that are dominant
in one area may not be dominant in the future, so multiple copies
of the same object shifts in terms of powerpoints within a dynamic
environment. This means that the master node is represented from
one copy of the object to another copy of the object as the robot
learns more.
[0343] FIG. 46C-46F demonstrates that the master node of B2 can
be represented from different copies of B2 in memory.
[0344] (FIG. 46C) On day1 the most dominant copy of B2 is on the
Beach with 11 points. (FIG. 46D) Then on day2 the library B2 and
supermarket B2 merged into one copy and became the dominant copy
of B2. (FIG. 46E) Then on day3 the robot encountered B2 at the capital
and a copy of B2 is recorded there. (FIG. 46F) On the fourth day
both copies of B2 from the capital merges into one and it became
the dominant copy of B2 with a total points of 19.
[0345] The network will keep on storing and modifying information
in the network based on what it senses from the environment. The
most important data that are trained often are kept in the network
and data that don't get trained often gets deleted from the network.
This works for all data types (all 5 senses and hidden objects)
in memory including: individual objects, floaters, pathways, scenes,
and complex situations.
[0346] Self-Organizing of Entire Pathways and Situations
[0347] In the last section we explored how the AI can self-organize
individual objects like people. In this section I explore how self-organization
averages entire pathways or situations in memory. I will use the
ABC block problem again. This problem is widely known in computer
science and scientists have been using this example to demonstrate
AI techniques in software programs.
[0348] The AI program must learn how to solve the ABC block problem
from a teacher. Teacher in this case can be teachers in school,
parents, friends, or anyone that understand the ABC problem. The
robot will take in the movie scenes and store them in memory frame
by frame. The location that the ABC block problem was thought is
where the AI will store that movie scene. If the robot learned how
to solve the ABC block in school then the movie scene will be stored
in the school location in memory. If the robot learned how to solve
the ABC block at home then the movie scene will be stored in the
home location in memory. Where ever the robot encountered the problem
is where it will be stored in memory regardless of where the location
might be in memory.
[0349] FIGS. 47A-47B are flow diagrams depicting the process of
how newly created objects are trained in memory. The masternode
will keep track of all the same copies (or fuzzy copies) in memory.
If one copy is modified in terms of data or powerpts then the masternode
will send a signal to all (or most of) the copies in memory to modify
its internal data.
[0350] In the first diagram (FIG. 47A) the newly created R1 in
memory is stored in memory. Next it sends a signal to the masternode
identifying itself. The masternode will make a note on this and
change its own powerpoints. Then it will send signals to other copies
of R1 in memory to increase its powerpoints depending on the strength
it has with the master node (FIG. 47B). If the connection is weak
then the increase will be low. If the connection is strong then
the increase will be high. In FIG. 47B the masternode's powerpoints
has been increased from 40 to 45. The second strongest copy of R1
(besides the newly created R1) has powerpoints of 8. The masternode
increased the powerpts by 2 points. On the other hand, the copy
of R1 with 3 points had an increase of 1 point and the copy of R1
with 1 point hand no increase at all because the connection was
too weak.
[0351] This type of retraining of data in memory not only works
for R1 but also R1's encapsulated format. Since there are many encapsulated
objects within R1, the AI will train the encapsulated objects in
R1 based on priority--the most important encapsulated objects get
trained first before the least important encapsulated objects get
trained (priority of objects and getting the encapsulated format
are discussed in later sections). A certain time limit is given
to the AI to retrain, self-organize, and find patterns to data.
When that time limit is reached it will stop storing and modifying
data.
[0352] Self-Organizing Entire Situations
[0353] This part is a little tricky and is more complex than training
individual objects in memory. There are several points I want to
clarify first before moving on. The target object is stored along
with its activated element objects (FIG. 48A). If the activated
element object is equal to the target object then both are considered
the same exact object.
[0354] In FIG. 48B, the object R1 and Meaning1 are considered equal
and are not separate objects. As time passes R1 and its encapsulated
objects (indicated by capital letters) begin to forget and data
disappears. The same will happen to data in Meaning1. Usually the
meaning of a sentence remains strong while the sentence that relates
to the meaning is weak. This means that the meaning will stay in
memory and the sentence will disappear. When data in the meaning
begins to disappear it will become a partial data G1 (FIG. 49).
[0355] Referring to FIG. 49, if meaning1 in memory forgets partial
data G1 remains. When searching for data in memory the AI tries
to find the optimal pathway. In the example in FIG. 52 partial data
G1 was found to be the most optimal pathway to choose based on a
meaning of a sentence that is similar to the meaning of sentence
R1 (well, partial data of the meaning of sentence R1).
[0356] Target objects R1 and R2 are considered similar but not
equal (FIG. 51 and FIG. 52). In FIG. 52, instead of matching R2
with data in memory, the AI matched R2's meaning (Meaning2) with
partial data of meaning1 in memory. Like I said before R1 and Meaning1
are equal and R2 and Meaning2 are equal. In memory, R1 has been
forgotten so we can't try to match R2 with R1. However, the meaning
to R1 remains in memory and the meaning is what will be used to
match the data from the current pathway to the data in memory. In
the case of R2, the AI activated Meaning2 as the meaning to R2.
And because R2 and Meaning2 are equal we can use either one (or
both) to try and find the best match in memory.
[0357] Alternative Scenario:
[0358] The AI program will use the target object to match what
it is currently encountering first. Sometimes, both the target object
and the meaning are used to find data in memory at the same time.
If the target object can't be found in memory then it can use the
activated meaning to the target object to match data in memory.
The AI decides which pathway match is the strongest.
[0359] The AI program will use both the target object and the meaning
to find the best pathway match in memory. In the case of the target
object, visual text words and sound words can be deceiving because
different sentences, even with a slight variation can mean totally
different things. This is why the AI will take into consideration
both the target object and its meaning to make a decision which
pathway has higher points. FIGS. 52-53 are diagrams depicting how
the AI program matches pathways in memory. FIG. 53 is one example
of how the AI program decides which pathway in memory has the highest
match percent (Path1 is the optimal pathway). Notice that even though
the optimal pathway has a visual text match of 25 percent the AI
picked that pathway instead of Path3 where the visual text match
is at 90 percent. The meaning is more valuable and has more powerpoints
in terms of match percent and that is why the AI decided to pick
Path1 instead. The pathway in Path2 has its visual text forgotten
(the data is so distorted that it's unreadable). However, the meaning
still remains and that has higher match than Path3 where it has
visual text and a meaning.
[0360] Powerpoints of the pathway is also a factor in decision
making. The percentage match and the powerpoints of that pathway
are used in combination to find the best match. The diagram in FIG.
50 shows that the AI found a similar pathway to R7 (the current
pathway). The first pathway rank has its visual text forgotten so
zero percent for both the match and powerpts. On the other hand
the meaning has a match of 40 percent. Because the meaning was subject
to forget the original meaning (Meaning1) has been distorted. But
it has a very high pointpts of 98. On the other hand pathway rank
2 has a 72 percent match but the powerpts is very low with 5 pts.
The AI picked the pathway with a meaning of 40 percent match and
98 pts. This illustrates how powerpoints affect the way decisions
are made in the AI program.
[0361] The Averaging of the ABC Block Problem
[0362] In previous sections we discussed how to average individual
visual objects in memory such as people and items. In this section
I have extended the object to include entire situations. Imagine
that R3 represents the ABC block problem. If a child was thought
the ABC block problem at school, at home, and at a neighbor's house,
then how does the average of the ABC block problem look like in
memory? The answer is we average out the object just like how we
average out individual visual objects in previous sections.
[0363] The diagram in FIGS. 54A-54B shows how the average location
of the ABC block problem is created and stationed in memory. Imagine
that a child learned the ABC problem at school in two separate classrooms--classroom1
and classroom2. In classroom1 the teacher thought the child many
times in different areas of the room so the powerpoints is 50. In
classroom2 the teacher thought the child 2-3 times so its powerpoints
is 5. In two other areas the child was thought how to solve the
ABC block problem by parents or neighbors. In the neighbor's house
the neighbor thought the child how to solve the ABC block problem
2 times so the powerpoints is 4. At home the child was thought the
ABC block problem 4 times by his parents so the powerpoints is 6.
In FIG. 54A self-organization will knit R3 together and average
out the location it should be in (location points 166 and 168).
Referring to FIG. 54B, notice that R3 170 didn't move much from
classroom1. The reason is because the majority of training examples
came from classroom1 and the average copy of R3 is closer to classroom1
than classroom2. In the second diagram two copies of R3 remain.
One is located near classroom1 (R3 170) and the other copy is located
between neighbor's house and home (R3 172). Because the two copies
are so far apart they are not subject to self-organization.
[0364] If a new copy of R3 is created in memory, that copy will
send a signal to the masternode and the masternode will increase
the powerpts of every (most) R3 copies in memory depending on the
connection strength. So, regardless of where the ABC block problem
is encountered the AI program will train itself globally. If two
or more copies of R3 are located in the same general area the self-organization
function will knit those R3 copies together and free up disk space.
The masternode will also be reassigned if one of the copies in memory
besides the masternode has the highest powerpoints.
[0365] The storage of data would include both the target objects
and the element objects activated by the rules program (FIG. 51).
When new data is created (this includes the element objects activated
by the rules program), a copy of that created object must send a
signal to the masternode. "Both R1 and Meaning1 must sent a
signal to their respective masternode after that data is created".
This is how data in the network is trained globally.
[0366] Language Organizes All the Data in Memory
[0367] Language brings order to chaos in our world. Language is
used to classify things that we learned to be the same and this
is a valuable asset to intelligence. Extremely complex intelligence
needs a very sophisticated language in order to develop. Without
language complex intelligence can't develop.
The whole idea behind the human level artificial intelligence program
is to build a software that can learn language and using language
to organize all the data in memory.
[0368] FIG. 55 is a diagram depicting the organization of data
in memory based on learned language. Because language looks the
same visually (words, letters, strings of letters, and sentences),
they are already closely grouped together in memory. And because
we learn language generally in the same area by a teacher, it is
grouped even closely. From all the school that a human being has
gone through--grade school, intermediate school, high school, and
college, the knowledge acquired over the years was learned in classrooms
or televisions or computer monitors. Because we were stationed in
one area for a year to learn knowledge the computer was able to
organize those data adequately.
[0369] What I'm trying to say is that language is organized in
memory in terms of visual representations and sound representations
(visual words and sound words). All the meaning to language is also
established in memory in one general area. The whole language database
is the organizer the AI uses to classify all data coming into memory
regardless of what sense it came from--sight, sound, taste, touch,
and smell (block 174). If new sensations are encountered the computer
will know where to organize that new sense in memory. If similar
data is sensed it will organize that sense in the most appropriate
area in memory. In other words the learned groups organize the data
in memory (block 176). The self-organization organizes both the
learned groups and commonality groups. Thus, giving the network
the power to learn language and use language to organize data in
memory (FIG. 55).
[0370] Hidden Data
[0371] Human conscious thoughts doesn't just have one function
it serves, but it does many things at the same time. As always language
is what organizes these thoughts. Language can tell us what the
meaning to words/sentences are, it can tell us information about
an object, or instruct us to solve complex problems.
[0372] In previous sections I discuss the 7 stages of how human
intelligence is developed. These 7 stages include a lot of things
such as learning the meaning to words/sentences, learning to plan
tasks, solve problems, copy other peoples' behavior and so forth.
All these things are leading up to one thing and that is to understand
and learn all the meaning to most words/sentences in the English
language.
[0373] When that understanding of every word/sentence is established
then we can use the self-organization function to encapsulate entire
situations in terms of language. Understanding words/sentences means
finding the meaning to words/sentences by finding the complex patterns.
Solving the ABC block problem is one example that I have used to
demonstrate how language is so crucial to learning ambiguous situations.
All the steps to solving the problem come from sentences. The movie
sequences that all training examples have do not look similar in
any shape or form. The sentences used (most notably the meaning
of the sentences) is what binds all the training examples together.
[0374] In this section I will explore the different ways that conscious
thoughts produce intelligence in humans by giving examples. Some
of these examples have already been used many times in this patent
but it is necessary to understanding how complex intelligence is
formed.
[0375] What kinds of data or functions are used to find complex
patterns to language?
[0376] In visual frames there are hidden data set up by the programmer
that will provide additional information about a movie sequence.
These hidden data are set up to establish additional data and allow
the AI program to find patterns that can't be recognized by what
is actually on the visual frames. Action words such as jump, walk,
throw, and run have patterns that can be identified by these hidden
data. Also, patterned sentences from hidden data can provide meaning
to object interaction. Below demonstrate patterned sentences. R1,
R2, R3 can be any object.
[0377] 1. R1 is on R2.
[0378] 2. R1 is walking toward R2.
[0379] 3. R2 is on R3 and R3 is on R1.
[0380] 4. go around R1.
[0381] 5. R1 is 3 feet from R2.
[0382] 6. R1 is below R2.
[0383] 7. R1 is under R2 but over R3.
[0384] 8. R1 collided with R2.
[0385] The hidden data is wired to the visual frames. All the image
layers or what is considered an image object will have measurements
that provide the AI with information about where that image object
is in relations to other image objects in the movie frames. The
hidden data also provide information about the properties of the
image layer such as the center point of the image layer and the
overall pixel count.
[0386] Since the hidden data is wired to the visual frames that
means the learned group that is equal to the visual frames has a
reference to the hidden data. This is important because the AI will
use a combination of the three groups in order to find complex patterns
and assign these complex patterns to sentences.
[0387] A note on hidden data, when the visual image (commonality
group) is forgotten, the hidden data still has the learned group.
If both the commonality group and the learned group are forgotten
then the hidden data stands alone. "The hidden data can exist
without either a learned group or a commonality group or both".
[0388] Hidden data contained in the visual frames:
[0389] For different senses the hidden data are represented differently.
For simplicity purposes hidden data from visual movies will be discussed.
These are the hidden data for visual movies: [0390] 1. Each image
layer has a normalization point (center point for that image). [0391]
2. Each image layer has a location point in the frame. The point
is the normalization point. [0392] 3. Each image layer has an overall
pixel count. [0393] 4. Each image layer has data that summarizes
all the pixels that it occupies including pixel color, neutral pixel
count, patterns in the pixels and so forth. Image layer (or image
object) interaction from frame to frame: [0394] 1. Each image layer
will have a direction of movement (north, south, east, west, northeast,
southwest etc.). This can represent words such as north, south,
east, direction, down, up, bottom etc. [0395] 2. Each image layer
will have coordinate movement in terms of x and y from frame to
frame. This can represent words like: moving, walking slowly, fast,
slow, one step, stationary, taking a break and so forth. If this
data is combined with the direction of movement then more words
can be represented such as: moving south, jump, walk, throw, trajectory,
the car took a nose dive into the water, the book fell, turn around,
jump up, look down, move sideways and so forth. [0396] 3. Each image
layer will have relationships to other image layers in the current
pathway. The relationships will include the coordinate points between
the two image layers and the direction between the two image layers.
[0397] 4. Each image layer will have a touch sensor that lights
up when it touches another image layer. This can represent words
like: touch, collision, slide, skim, and so forth. [0398] 5. Each
image layer will have a degree of change from one frame to the next.
If it changes its shape dramatically it will be recorded. If it
changes its shape gradually it will be recorded. This is important
because if the image layer touches another image layer the degree
of change will tell if the interaction caused the image object to
change or it didn't cause the image object to change. A car accident
definitely changes the way a car looks after the collision, while
solid objects moving very slowly and colliding don't change its
shape. [0399] 6. Each image layer will have scaling and rotation
data. Did the image layer grow larger in size? Did the image layer
rotate to the right? If it did what is the degree of rotation? Words
such as: grow bigger, deflated, change its size, rotated, towards,
move away from, and shrink can be represented by this data.
[0400] These are just some of the hidden data that will accompany
visual images and movie sequences. The programmer can add in more
data, but the AI will take a longer time to find patterns among
the hidden data. This is where the programmer should decide how
much hidden data to include. Too much hidden data will overwhelm
the system and too little will prevent the pattern function from
doing its job properly.
[0401] FIGS. 56A-56B are diagrams demonstrating the 3 types of
data in the current pathway: 5 sense data, activated element objects
and hidden data. The diagram in FIG. 56B is the same diagram in
FIG. 48B but I included the hidden data in the current 5 sense pathway.
All the visual images in the current pathway will be broken up into
image layers and determined their respective 360 degree floaters.
Each image layer generate hidden data and establish relationships
to other image layers in the movie sequence. R1 is stored in memory
along with its hidden data (FIG. 56A). Then the rules program activates
Meaning1 based on the target object R1. This means R1 and Meaning1
is the same object. This also means that the hidden data located
in R1 is shared with Meaning1 (FIG. 56B). If the AI program forgets
R1 in memory and the hidden data hasn't been forgotten then Meaning1
will have the remaining information from the hidden data (hidden
data is subject to forget as well).
[0402] All three groups: commonality groups, learned groups and
hidden data are subject to forget.
[0403] Forgetting in Commonality Groups
[0404] FIGS. 57A-57B are flow diagrams illustrating how commonality
groups or 5 sense data forget information. Commonality groups forget
based on what encapsulated groups are trained the most. If different
eyes are trained such as human eyes, anime eyes, cartoon eyes, dog
eyes and so forth, the eyes that are trained the most (the robot
encounter the most) will be dominant. Another example are lines,
if the robot encounters a straight line more than a curve line then
the straight line will be dominant and will be a stronger object
than a curve line.
[0405] In FIG. 57A, the visual movie sequences (commonality groups)
will be stored in memory with DVD quality. As the AI forgets the
information (based on strength of commonality groups) the AI will
have its video quality lowered (FIG. 57B). By the time the information
is forgotten the movie quality is so distorted it is not recognizable
and the movie sequence is not connected anymore but broken up into
multiple sub-movies. Only the strongest memories get remembered
while the minor things get deleted.
[0406] Forgetting in Learned Groups
[0407] FIGS. 58A-58D are diagrams illustrating how learned groups
or activated element objects forget information. The learned groups
forget information in terms of objects encapsulated in that learned
group. The sub-learned groups leading to that learned group is used
to degrade information. Imagine that you looked at a leg of a horse
then you moved to the neck of the horse then to the head of the
horse. The learned groups leading to the activated word "horse"
is presented this way: leg.fwdarw.neck.fwdarw.head.fwdarw.horse
(FIG. 58A-58B).
[0408] Humans see things not in terms of frames in movies where
the pixels are equal in visibility (FIG. 58C). The human eye focuses
on an image. The image it is focused on is clear while images that
fall in its peripheral vision are blurry (pointers 180 and 182).
[0409] In the example in FIG. 58B the robot focused on the leg
first, then it moved to the neck, and finally it moved to the head.
It is at the point when the robot identified the head when it recognized
that the image layer is actually a horse (FIG. 58D). Some people
would see the leg, and because they are experts, they identified
that image as a horse. For most of us when we see the leg we might
think it's a donkey or a dog. For different people identification
and activation of image objects is different.
[0410] Since the leg, the neck and the head is part of the image
layer and that image layer is identified as a "horse",
then leg, neck, head are all objects encapsulated in the sound "horse".
The AI will store this data in memory and the encapsulated objects
will forget based on its encapsulated format (FIG. 60). Whichever
objects are the strongest gets forgotten last and which objects
are weak (low powerpts) will be forgotten first.
[0411] The learned groups have coordinate points (From the hidden
data of the visual image layer the learned group equal) on each
frame. The objects that are contained in a learned group will be
considered its' encapsulated objects. This is why leg, neck and
head are all learned groups contained in the "horse" learned
group. Each of these learned groups will correspond with the normalization
point that their visual image layer has. That is why the leg group
is below the neck and the head is to the left of the neck (FIG.
60). The "horse" group encases the leg, neck and head
and has a normalized point that is the center of the leg, neck and
head. Also, the AI need not activate the word for that image layer.
For example, if the image of the leg is encountered by the AI the
sound "leg" may not activate, instead maybe a reference
of the leg floater is activated or something else that is equivalent
to an image of a leg.
[0412] Another theory is that the AI uses strong learned groups
"in memory" to forget information. All the strongest sub-learned
groups contained in the learned group will be used to forget information.
The strongest sub-learned groups will remain and weaker sub-learned
groups will forget. It could also be both theories above that learned
groups forget information.
[0413] Forgetting in Hidden Data
[0414] FIG. 59 is a flow diagram illustrating how hidden data forget
information. Data in hidden data are called elements. Each element
doesn't have a predefined priority. Instead, the priorities of the
elements depend on pattern groupings and pain/pleasure. If the AI
encounters certain elements over and over again, it will have a
higher priority number (common groups having the same elements and
grouped together). If the AI doesn't encounter certain elements
and that element isn't trained often, then that element will have
a lower priority number (common groups don't have this element).
Another factor to priority of elements is pain and pleasure. Pain
and pleasure is discussed in other parts of this patent but I will
summarize what it does. When the robot encounters pain all the pathways
and its encapsulated format leading to that pain will have their
powerpts decreased. On the other hand when the robot encounters
pleasure all the pathways and its encapsulated format leading to
that pleasure will have their powerpts increased. Objects in the
pathway closest to the pain/pleasure will have their powerpts modified
strongly while objects farther away from the pain/pleasure are modified
mildly. The AI program tries to locate the object or objects that
caused the pain/pleasure. When the AI program identifies the objects
that caused the pain/pleasure then it will assign higher priority
to those objects.
[0415] What are the hidden objects assigned to words/sentences?
[0416] In previous sections we talked about how words/sentences
are assigned meaning using the rules program. The meaning of the
words/sentences is actually hidden object or a combination of hidden
data, commonality groups, and learned groups all combined together
to form a complex pattern (in this section a fourth group is discussed,
patterns). This meaning (complex patterns) is then assigned to something
that is fixed. Since language is fixed the rules program assign
the meaning to words/sentences.
[0417] FIGS. 61A-61B are diagrams illustrating how the AI program
reads in the word bat. When the robot reads text from a book it
reads text exactly like a human being. From a movie sequence the
words are seen one letter at a time. The letters are focused on
and identified by the robot. The recognizing of these sequential
letters make up words that mean something. FIG. 61A is an example
of how the robot will identify the word "bat".
[0418] At this point, while we read in each letter of the word
the sound that accompanies the letters are pronounced in the mind
(FIG. 61B). That sound is the meaning because it has very strong
associations with the letters. By the time the robot finish reading
in the "T" the sound "bat" will pop up in memory.
At that moment more element objects that have association to the
visual text "bat" activates in the mind--element objects
such as a picture of a bat.
[0419] Small length words such as "bat" can be identified
in memory without reading in every single letter. The whole text
image "bat" can represent the word. But much longer length
words like "computerization" might require the robot to
focus and identify multiple sequential words in order to understand.
(since there are so much meaning to the word "bat" the
conscious will tell the robot what type of bat it is. If reading
a book, there are other words and suggestions to indicate what type
of bat the word means. The lessons thought in English classes will
guide the robot to look for clues here and there to find the true
meaning to the word "bat").
[0420] The movie sequence of recognizing words/sentences is actually
stored in terms of fuzzy logic. The text in the movie sequence can
be in any font or it can be in any font size. The paper can be in
any color or the text can be on a computer screen or a wall. You
can even line up chopsticks to represent the text. The different
ways of expressing the word "bat" can be infinite but
the meaning to the word will always be fixed. When the AI program
averages out all the training examples a fuzzy range of the movie
sequence will be created in memory. In this moment the different
ways of expressing the text word "bat" can be understood
by the AI program regardless of how distorted or fuzzy that movie
sequence may be. But there is a threshold in which different movie
sequences will be considered the word bat or not. (the meaning of
the words/sentences will also be in a fuzzy logic way. In fact,
all data in memory will be in a fuzzy logic way. This is the whole
point about building a network that can store infinite data).
[0421] All Objects Created in Memory has a Default Learned Group
[0422] When data is created in memory it will automatically be
assigned a default learned group called default object. Anything
that it is assigned in the future will be derived from default object.
For example, if the robot learns one cat image that wasn't learned
before, the robot will store this newly created image and assign
it a default object. When the robot learns more and a floater is
created of this cat object the rules program will assign the learned
group "animal" to the floater. This means the sound "animal"
and the 360 degree floater of cat is equal. The learned group "animal"
is derived from the default object. Now, we can give the cat floater
a more specific identification. We can train it to identify the
360 degree images of cat and assign it to the learned group "cat".
Although animal is one possible learned group to identify the cat
images, the learned group "cat" is a more specific term
used to represent the 360 degree images of a cat. This encapsulation
of learned groups to identify an object is created in memory. The
AI program will activate the most specific learned group to represent
an object. In this case the most specific learned group is the sound
"cat" to represent the visual images of a cat.
[0423] FIG. 62 show different learned groups assigned to the same
360 degree floater of cat. The most specific learned group has the
strongest connection. In this case the sound "cat" has
the strongest connection weight to the 360 degree floater of cat.
All objects in memory regardless of how weak is referenced to a
default object. You will see later how these encapsulated learned
groups will be used to find meaning to sentences.
[0424] Hidden Objects
[0425] The whole point about hidden objects is that only the computer
knows what these hidden data are. The computer will take the data
from the current pathway and average this data out with the data
in memory. The data in the current pathway have four types:
1. data from the 5 senses (commonality groups)
2. data activated by the rules program based on the 5 senses (learned
groups)
3. hidden data embedded in the 5 senses.
4. patterns and identification
[0426] In the previous section I added hidden data to the 5 senses
and explained how that data is integrated with the current pathway.
In this section I have included one more data type (patterns). FIGS.
63A-63B are diagrams demonstrating the 4 types of data in the current
pathway: 5 sense data, activated element objects, hidden data and
patterns. Patterns are created in the current pathway only after
self-organization. This part is very important to convey fuzzy logic
in pathways and how patterns are used to create universal pathways.
(I slowly introduce these data types so that the reader can understand
what the current pathway contains and to understand each data type
thoroughly).
[0427] In terms of searching for data the AI program will use the
common traits of the current pathway (the first 3 data types) and
compare them with the data types in other pathways in memory. In
terms of self-organization the AI program will group common traits
together and either create or strengthen existing common traits
in memory. It also finds patterns within these common traits and
creates a sort of patterned sequence based on the four data types
above. After the self-organization is done the most dominant hidden
objects in memory will stand out from the weaker hidden objects.
The hidden objects will be assigned to words/sentences by the rules
program and will represent meaning to language.
[0428] In some sense only complex words or sentences have hidden
objects as meaning. Other data from the 5 senses are much more straight-forward.
For example an image of a dog has the meaning of the sound "dog",
the sound of a cow `mooing` has the meaning of a visual image cow,
and the visual image of a cat has the meaning of the sound "cat".
These are simple examples of meaning to words. A more complex form
of this is putting these words together to form sentences. This
form of sentence objects need a more complex way of stringing meaning
together and that is why hidden objects are used to assign meaning
to complex words or sentences. A word like "universe"
isn't something that can be represented with a visual image (it
can be). But a true meaning of the word has to be from complex intelligence
and this complex intelligence can only be formed by using hidden
data and finding and fabricated patterns within these hidden data.
[0429] Patterns and Identification of 5 Sense Objects
[0430] All four types of data in the current pathway will be used
to find any repeated patterns with similar pathways in memory. These
four data types are: commonality groups, learned groups, hidden
data, and patterns. Instead of explaining what the different kinds
of patterns exist I will use simple examples to illustrate this
point.
[0431] 1. Sentence Represented by Sound
[0432] The first thing the AI program will do is identify the 5
sense objects from the current pathway. If there was a picture of
the text word "cat" and right next to the text word is
a picture of cat, then we identify that the text word "cat"
is identified by the visual picture of cat. If someone said "cat"
and pointed at a visual image of a cat then that sound "cat"
identifies the visual picture in the movie sequence. (a more complex
way of identifying 5 sense objects is through conscious thought.
This will be discussed in later parts of the patent).
[0433] If the AI program can't identify the 5 sense objects from
the current pathway then it will identify the 5 sense objects using
the default way--identified in memory. For example, if the robot
had no visual sight and the only sense his got is sound, then when
the sound "cat" is recognized by the robot, the identification
is referred to in memory. If the robot has sight and sound then
if a cat image is within the robot's sight then the sound "cat"
is identified as the visual cat image.
[0434] Now, imagine there were two sentences, one with a question
and another with an answer. The robot only have one sense: sound.
These sentences are sound recognized by the robot. Since there is
no vision the AI program will refer to the data in memory.
Sentences: What is 5+5? 5+5 is 10.
[0435] Identification of the words/letters in the sentence happens
sequentially. The words/letters are known as the target object.
The AI searches for these target objects and activate element objects
that have strong association to the target objects (learned groups).
The AI program will attempt to use measurements in the hidden data
to find patterns. Since sound is linear data we don't have to worry
about 3-dimensional space. But time is important. "The computer
will average out the timing of similar pathways".
[0436] FIGS. 64A-64C are flow diagrams showing how the AI program
finds patterns to similar pathways and output a universal pathway.
FIG. 64A is an example of one pattern found. Equal objects is very
important. The AI will attempt to establish an equal connection
with the sequential data 186 in the current pathway and data 184
from memory. Once all these data are found, the AI will compare
this example pattern with other similar examples in memory and establish
a universal pathway. This universal pathway contain the instructions
to find future data based on the current state (FIG. 64C). For example,
if the AI encounters: What is 5+5? and the current state is at the
end of the question, then the future prediction has already been
established based on the pattern in the example (FIG. 64A-64B).
The future prediction is "5+5 is 10". Other similar Q
and A can be predicted such as:
What is 8+8? 8+8 is 16.
[0437] The AI program averages out the patterns at every sequence
in the pathway so that regardless of what state in the sequence
the AI is in it already has a copy of the pattern that it needs
to predict the future.
[0438] In FIG. 64C the universal pathway will contain the average
of all similar examples in memory. Data 188 found in memory and
data 190 from the 5 senses have patterns and the patterns are indicated
by dotted arrows in the diagram. The default learned groups will
accompany the data (target objects). The example shows that E1 and
E2 are represented by default learned groups and this default learned
group can be anything.
[0439] The learned groups that accompany the target objects may
not be a default object but any of its hierarchical learned groups.
The example in FIG. 65A show using a cat and a dog as target objects.
Although they are different the fact that they share a hierarchical
learned group establishes an equal pattern. The sentences don't
make any sense but you got the point.
[0440] Both cat and dog are animals so that learned group will
accompany the pattern to find more specific types of data (FIG.
65B).
[0441] The examples in FIGS. 65A-65B show that using hierarchical
learned groups that are shared among data can lead to a more defined
and specific pattern. The more specific the pattern is the better
the future prediction is.
[0442] The timing of when the target objects (words) occur is averaged
out by the AI program and a fuzzy range of how the sequences occur
will be added to memory. The closer the timing of the target objects
the more accurate the future prediction will be. Also the length
of the target object is also averaged out so a word like "computerization"
and a word like "bat" can be represented as the same object
in the pattern. Remember we are only dealing with sound here (sound
words). These words and sentences are linear in order. In the next
section we will discuss how words and sentences are interpreted
on a 3-dimensional visual environment (visual text words).
[0443] FIG. 66 is a diagram showing the different times events
occur. Actually the timing of when objects occurred is part of the
hidden data attached to the current pathway. The time of S1, S2,
and S3 target objects recognized at different times. The AI program
averages out the time and output a universal pathway that give an
approximate time certain objects occurred.
[0444] 2. Sentences Represented by Visual Text
[0445] When dealing with language on a visual 3-dimensional space,
the AI has to worry about position of the letters/words. Text words
on books and monitors are language represented on a visual 3-dimensional
space. Just like sound words/sentences the AI program identifies
words/sentences sequentially. This time the position of the words
is a factor that must be taken into consideration.
[0446] FIG. 67 is an example of two similar sentences but in sentence
B the word box is not centered as sentence A. These two are not
considered identical, even though the computer reads in the words
in the same sequential manner.
[0447] Using visual means to represent language is far more advance
and has a lot more capabilities than representing language with
sound. For one thing we can now manipulate visual images on the
frames by moving images, deleting images, creating images, identifying
images and assigning one image to another image. Language can now
be represented in such a manner that the possibilities are limitless.
[0448] There is no such thing as built in assignment statements.
In my AI program objects are assigned to other objects in terms
of activation by the rules program. If the text word "cat"
is encountered and the rules program activated an image of cat,
then the cat image is equal to the text word "cat". The
only way for this to happen is if the text word "cat"
is encountered many times along with the visual image of cat. Both
the text word "cat" and the cat image is strong in terms
of association that they are considered equal. The example below
demonstrates this idea.
[0449] FIG. 68 is an illustration of a mouse and the text word
mouse. If you keep showing a mouse picture and the text word mouse
then these two objects will have greater and greater association
to each other. When the association between the two are strong enough
one object will activate the other object and vice versa. The example
in FIG. 69 shows what happens when the text words mouse is identified
by the AI program. The visual picture of mouse gets activated. The
next time that we see the visual text mouse and a visual picture
mouse, the AI program will identify that mouse picture with that
text word mouse.
[0450] FIG. 70 is an illustration of how the AI program identifies
the word mouse in the movie sequences. In the movie sequence when
the text word mouse is identified, the AI program assigns this mouse
word to the mouse picture in the next frame. The AI could have assigned
the mouse word to the cheese picture but these two objects aren't
equal. This technique is also used in words that take up multiple
sequential frames in a movie--a word like jump.
[0451] FIGS. 71A-71B is an illustration of how the AI program assigns
the word jump to a movie sequence. The jump word is assigned to
the jump sequence of the dog. The cheese is not part of the word
jump (FIG. 71B).
[0452] These simple examples are used to demonstrate that when
a sentence is identified by the AI program it will also identify
if words in the sentence have a reference in the movie sequence.
A more complex example is the sentence: the dog jumped over the
box. The AI will try to find all the objects that are involved in
the sentence. The object dog is involved. That means all the sequential
image of dog will be cut out from the movie. The jump sequence is
involved so the sequence of the dog jumping is cut out from the
movie. The box is involved so the sequential images of the box will
be cut out of the movie. Using all these objects from the movie
the AI can combine these layers of the movie and form a sequence
that only involves the sentence: the dog jumped over the box. Patterns
are also involved to understand the sentence fully. For example
"jumped over" means that the dog image layer is positioned
above the box image layer.
[0453] Identifying Meaning to Sentences in the Current Pathway
[0454] The above example illustrate an exact meaning (the sequence
that reflect the sentence) to a sentence (the dog jumped over the
box). In real life the brain can only store a fuzzy range of the
meaning to a sentence and not the exact meaning. The self-organization
will average out similar examples in memory and forge a universal
sentence pathway to cater to infinite possibilities. This universal
sentence has a broader meaning that can cater to the example above
and anything that is similar.
[0455] I will explain the self-organization part further because
that will demonstrate how the universal pathway is created. If you
had three sentences such as:
[0456] 1. the dog jumped over the box
[0457] 2. the cat jumped over the box
[0458] 3. the rat jumped over the box
[0459] The meaning to this is quite apparent, simple replace R1
(default object) in the position in the sentence that has many variations.
This will create a pattern in which during runtime the AI can replace
R1 with the appropriate object and the meaning can be understood.
Universal sentence: the R1 jumped over the box.
[0460] The sentence can be even more universal by averaging the
other object in the sentence: R1 jumped over R2. Now, the AI finds
the meaning by replacing R1 and R2 with its appropriate objects
during runtime. That fabricated sequence is the meaning to the sentence.
[0461] In this section the topic is: identifying meaning to sentences
in the current pathway. This means that we have to identify the
elements and patterns in the meaning and try to find the sequence
that it belongs to in the current pathway. At this point the sentence
that activated the meaning has nothing to do with this. Once the
rules program activated a meaning to a sentence, that meaning has
to be identified either in the current 5 sense pathway or in memory.
(remember I said that all objects, target objects or element objects,
must be identified). This is important because self-organization
will group different or similar sentences together that have similar
or the same meaning.
[0462] FIG. 72 is a diagram of different sentences assigned to
the same meaning. Although all the sentences are different the meaning
is virtually the same. This groups all the different sentences together.
This is how the AI program will understand the same meaning of a
situation regardless of what sentence is being used to explain the
situation. The AI program will use all the elements from the meaning
(Meaning5 192) and try to identify the sequence of the sentence
from the current pathway. If there is no sequence from the current
pathway that matches the meaning then it will be assigned the default
setting which is identification of meaning in memory (for example,
if the sentence was sound and the robot closes its eyes no sequence
will be identified in the current pathway. But images and movie
sequences from memory will activate providing a fabricated movie
sequence).
[0463] More Examples of Fuzzy Logic
[0464] In this and the next section we will discuss about fuzzy
logic and how sentences are represented in terms of fuzzy logic.
Things that we say in a language can mean the same things. Sentences
such as:
1. "stack up the blocks in an A B C format"
2. "I want you to stack the blocks up starting with C then
B and finally A"
3. "can you please stack the blocks up in alphabetical order"
[0465] Although visually the sentences look different they mean
roughly the same things. The meaning is what brings these three
sentences together. This is what I mean by representing words/sentences
in terms of fuzzy logic. The three sentences above are said during
a particular situation but the exact sentence is not encountered
everytime. This will store the sentence temporarily in memory because
it doesn't repeat itself, while on the other hand the meaning and
its encapsulated formats become stronger and stronger.
[0466] In this section I will try to present simple examples to
illustrate my point about how meaning is assigned to words/sentences.
Some of these examples might contradict my previous lessons but
let's just say that there are several ways of accomplishing the
same things.
[0467] The next example is to find the meaning to sentences that
have this structure:
"F1 on F2"
[0468] F1 and F2 are variables assigned at runtime and it could
be anything. Among some of the sentences that fall into this category
are: triangle on square, circle on square, square on pentagon, mouse
on cheese, and so forth. The sentence structure "F1 on F2"
is a universal sentence that will cater to infinite possibilities.
[0469] In order to create this universal sentence the meaning of
all the sentences have to be the same or similar. The variables
F1 and F2 are default objects (default learned group assigned to
a 5 sense data). Since all objects in memory are derived from a
default learned group then F1 can represent any object in memory.
[0470] In example1 a triangle is on top of a square (FIG. 73A).
On example2 is the learned groups and the hidden data that accompanies
all the image layers (FIG. 73A). The dot in the center of the triangle
and the center of the square is the normalization points. It is
accompanied by the coordinate point in the frame. The learned group
for each image layer is also attached to the image layer. The triangle
image is accompanied by the learned group "triangle" and
the square image is accompanied by the learned group "square".
In FIG. 73B the contact point 198 between the two image layers is
shown. These are just some of the hidden data that is attached to
the image layers, there are many more.
[0471] When the AI finish assigning these hidden data and learned
groups to the image layers it will then establish relationships
between image layers. FIG. 73C is an example of some of these relationships.
The triangle is north west in relation to the square. The square
is south east in relation to the triangle. The triangle is in contact
with the square which means it is touching one another. The contact
location is delineated by the dotted line.
[0472] Referring to FIG. 73D, after averaging out similar pathways
in memory the computer will have a universal meaning that all examples
have (or at least the majority of the examples have). Pointer location
200 is the different variation of the "triangle on square"
examples. All three examples contain the data in Meaning6 (block
202).
[0473] In all three examples the meaning is the same. All statements
in Meaning6 are true for all three examples. In fact, you can come
up with infinite variations to visual images of a triangle on a
square and the computer will still generate the same meaning.
[0474] In terms of what sentences are assigned to what movie sequences,
it will all depend on the rules program to find the association
between two objects. The more times you train a sentence with the
movie sequence the stronger that association will become. The closer
the timing of the sentence with the movie sequence the stronger
the association will become. This means that the meaning can be
assigned to any sentence and the meaning can be changed. For example,
I can assign "triangle fly square" to Meaning6. All I
need to do is train the rules program so that "triangle fly
square" is assigned to Meaning6. I have to train it so that
this sentence overpowers the previous sentence: triangle on square.
All of this would mean that I can use words/sentences from different
languages to represent the same meaning. This is why this form of
language learning is universal.
[0475] Extension of the Last Example
[0476] Now we add in the learned groups to this example and see
how a universal pathway can be applied to "F1 on F2".
FIG. 73C are 3 similar examples of "F1 on F2". The only
real difference is that the image layers are different, but the
meaning of the sentences is the same. In order to create a universal
meaning (FIG. 73E) for these examples we have to replace the image
layers with their respective learned groups that all examples have.
In this case all three examples have default object, so that will
be the learned group that will represent the meaning.
[0477] The image layers can be represented by other learned groups
as well. However, all three examples above share only one learned
group--the default object. On the other hand, if we learned that
all image layers in the examples are shapes. Then we can replace
the default object with the learned group "shape". If
the image layers were animals like cat, dog, and mouse, we can use
the learned group: "animal" as the universal variable.
The more specific the learned group is the more specific the actual
movie sequence can be. The less specific the learned group is the
broader the movie sequence can be. In some sense all the pathways
in memory are hierarchical in nature and it goes from general to
specific. The AI program will most likely pick the most specific
to predict the future because the more specific the meaning the
more accurate the future prediction is the less specific the more
inaccurate the future prediction is.
[0478] This is why it is so important for the AI program to encounter
many examples of a situation in order to predict the future when
that similar situation is encountered.
[0479] Complex Sentences and Meaning
[0480] I would like to say that representing all words/sentences
in a language can be done by the method presented above but that
isn't how complex intelligence is created. In order to learn a complex
sentence, grammar rules are included in the sentence to understand
what different words mean and how the words interact with pictures
or images in our environment (FIG. 74). This is where the human
conscious comes in. Trees are instructions activated by the rules
program to instruct the AI program to understand meaning, give information
about an object or situation, or solve a problem. These trees are
usually in the form of sentences or visual (and sound) movies that
tell the AI what to do next.
[0481] Notice how complex understanding a sentence like the example
in FIG. 74 really is. Understanding a sentence comes from teachers
in school that thought you the rules of grammar in a particular
language. We use other words/sentences to encapsulate those lessons
and import these lessons to understanding structures and meaning
in sentences. The computer uses all the activated element objects
and the target objects to assign variables in meanings (meaning7)
to get a better idea of what all the words in the sentence really
mean.
[0482] Assignment Statement Example
[0483] In previous lessons I stated that assignment statements
are done by what is activated by the rules program from the target
object. However, in order to learn that the activation of element
object from target object is equal, we have to use patterns. The
sentence: "this is a mouse", require that a pattern is
found to state that the sentence is saying that an image in the
picture is the equivalent to the word "mouse". One pattern
that can be used is the equality of two objects. If two objects
are stationed in the same assign threshold they are considered equal.
In order to understand the sentence: "this is a mouse",
the AI program must find "this" and "mouse"
to be equal objects. In pattern finding the AI has to work with
all the patterns that the programmer has set up. I won't disclose
all the patterns just yet, but one of these patterns is the assignment
state or equality of two objects. The only way to find out that
two objects are equal is by looking at their respective location
in memory. Do both objects fall into the same assign threshold?
If the answer is yes then both objects are equal.
[0484] Referring to FIG. 75, in frame 208 "this" is referring
to the image the finger is pointing to. The image layer is a picture
of a mouse. In the sentence contains the words mouse and in the
frame contains an image of mouse. The pattern resides in data 210
in memory where the sound "mouse" and the image mouse
are equal. By averaging this example with other similar examples
the AI program will understand that the sentence "this is a
R1" is actually an assignment statement (FIG. 76A).
[0485] In FIG. 76A the AI program has to learn the different variations
of the situation from 360 degrees (Ex. 1 and Ex. 2). The finger
can be anywhere in the frame and the mouse can be anywhere in the
frame. Each image layer can be different but belong to the same
object. For example, the hand can be any image from the hand floater.
The mouse can be any image from the mouse floater. The sound "this
is a mouse" and the learned groups in the frames binds them
all together and the computer will find the common traits among
all the different examples.
[0486] We can extend the last example by introducing variables
in terms of the objects. The different ways of presenting the sentence
is illustrated in FIG. 76B.
[0487] The self-organization function will average the three examples
in FIG. 76B and create a universal pathway 212 that will cater to
all similar examples. This universal pathway 212 will be used to
understand the sentence the next time the AI encounters a similar
situation. The default object will be assigned an image layer in
the frame at runtime, the "finger" will be assigned an
image layer in the frame at runtime, and R1 will be assigned a sound
(word) at runtime.
[0488] Patterns
[0489] In the last section I have given one example of the assignment
statement. The assignment statement is one internal function used
to find patterns. The sentence "this is a mouse" demonstrate
how language represent the assignment statement (FIG. 75). In previous
sections I outlined another internal function to find patterns which
is searching for a particular data in memory and extracting information
from this data. Answering questions such as "What is 5+5? 5+5
is 10" is one example of how the AI uses internal functions
instructed by patterns in a pathway to predict the future (FIGS.
64A-64C).
[0490] I wanted to slowly introduce the different types of internal
functions that are available to the AI program to find complex patterns
within similar pathways. In this section I will outline most of
the internal functions that are used by the AI program and give
examples of these patterns. As always, words/sentences are used
to express how the patterns work.
[0491] Equal Objects and Hierarchical Learned Groups Establish
the Elements Involved in the Pattern
[0492] Equal objects and its hierarchical learned groups is what
will establish the data that we want to find patterns to. It provides
us with the means of sorting out what data are involved in the patterns.
Let's review on the question and answer example, "What is 5+5?
5+5 is 10" (FIG. 77A). The equal objects in the pathway and
the pathways in memory establish what will happen in the future
before that future happens.
[0493] Referring to FIG. 77A, imagine that we are at the current
state, the objects that we encountered can be used to predict what
will happen in the future. Objects from the question are used to
find what will happen in the future. Equal objects from the future
and equal objects from the past are used to establish the patterns
to get a future prediction.
[0494] The equal objects established in the pathway aren't just
the objects we need to find patterns. We have to look for these
objects in memory and find out the relationships between all the
equal objects in the current pathway as well as the equal objects
in memory.
[0495] Referring to FIG. 77B, after the AI determine all the equal
objects in the current pathway and the pathways in memory (indicated
by dotted arrows), the AI will compare this current pathway with
other similar pathways in memory. The pattern is the result of common
traits among all the similar pathways. After averaging the data
the AI program will determine that in order to predict the future
from the current state, the AI must use some of the objects from
the question and search and look for some of these objects in memory
and extract certain data from memory. The AI will utilize internal
functions in order to accomplish these tasks.
[0496] FIGS. 77A-77B is just a review on how the AI program finds
patterns to predict the future. Here are most of the internal functions
used by the AI program to find meaning to language and predict the
future: [0497] 1. the assignment statement--the rules program determine
the assign threshold. If two objects pass the assign threshold that
means both objects are equal. Patterns are used to assign this function
to a sentence. [0498] 2. searching for data in memory--This function
searches for and extract specific data from memory by using patterns
that were found by similar examples. The AI program can extract
data from linear sound, it can extract data from 2-dimensional visual
movies, or any other 5 sense data. [0499] 3. determining the distance
of data in the 3-d environment--finding the distance between two
or more objects in memory based on patterns. [0500] 4. rewinding
and fast forwarding in long-term memory to find information--the
length of when certain situations happen and where it happened is
based on patterns. Information will also be extracted from the movie
sequences. [0501] 5. determining the strength and the weakness of
data in memory. How strong is one data compared to another data
and how the data changes during a time period depend greatly on
patterns. [0502] 6. a combination of all internal functions mentioned
above.
[0503] These are just some of the internal functions that are being
used by the AI program. The most important is searching for data
in memory. Most of the time this function will be used to find patterns.
Instead of explaining each internal function to the reader, I decided
to provide examples to illustrate how they are used.
EXAMPLES
[0504] A. The assignment statement--the example in FIG. 75, "this
is a mouse", explains how this function works. The AI program
creates an assignment statement to the sentence "this is a
mouse".
[0505] B. Searching for data in memory--the example in FIGS. 77A-77B,
"What is 5+5? 5+5 is 10", explains how this function works.
The AI program uses similar pathways to find a universal pattern
to answer the question. It not only searches for certain data from
the problem but it also extracts data from pathways in memory.
[0506] C. Determining the distance of data in the 3-d environment
(data in memory).
[0507] The pathways sensed by the robot will be stored in memory
in a 3-d environment. For these 2-d sequential frames the AI will
store them so that a 3-d environment is created. This 3-d environment
will be used to find information.
[0508] FIGS. 78A-79B are diagrams showing internal function: finding
data from the 3-d environment. The question "Where is the bathroom?"
is a question that require the robot to use the 3-d environment
to extract the location of objects. In this case bathroom is the
object. This sentence is derived from "where is the W1".
W1 is a variable representing an object. The AI encountered many
examples of similar questions and was able to create a universal
pathway. There are actually two ways that this question can be answered.
The example in FIGS. 78A-78B present the first way to solve this
problem the other way to solve this problem is by using trees to
instruct the AI program to answer the question. The first way to
solve this problem is by observing the sequential events that occurred
and see if there are any patterns involved.
[0509] The AI will establish the target objects found in memory.
Then it will attempt to find patterns between similar examples (FIG.
78A).
[0510] So, based on these two similar examples (FIG. 78A and FIG.
78B) the AI will forge a universal question and answer pathway.
Instead of using visual data in our environment to find the patterns
the AI uses the visual environment in memory to find these patterns.
The current location is one floor to the cafeteria so the robot
will not be able to see the cafeteria nor the elevator. Instead
the robot uses the learned knowledge of the structure of the building
to find out patterns to the situation.
[0511] The next time someone asks the question: "where is
the principal's office?". Because the robot understands the
pattern the robot can answer the question. It will identify its
current location. Then it will locate the principal's office in
memory. Finally, it will output the location based on a visual picture
of the two destinations (the current location and the principal's
office). Outputting the answer to the question might be an encapsulated
instruction in terms of knowing how to say things in English and
interpreting locations of two places. These two knowledge has been
learned before by teachers and is incorporated into the pattern
by trees (sentences).
[0512] One more example to illustrate how the 3-d environment can
be used to find patterns is determining the distance between two
places (FIGS. 79A-79B). If the question: "how far is it from
the supermarket to the library?", the answer to this question
would require the 3-d environment from memory.
[0513] The distances in the 3-d environment have already been assigned
to language. So a certain distance in the 3-d environment activates
certain words that represent the distance. In FIG. 79A the distance
from the supermarket to the library is interpreted as 1 mile. This
1 mile is part of the answer the robot needs to answer the question.
If the robot compares this example to other similar examples then
a pattern is found. The universal pathway is presented in FIG. 79B.
[0514] D. Rewinding and Fast Forwarding in Long-Term Memory to
Find Information
[0515] The next internal function is having the ability to rewind
or fast forward experiences the robot has encountered. All the movie
frames are stored in a timeline when it accord and the AI program
breaks up the movie frames into sections and store these sections
in memory. The long-term memory is this timeline and the timeline
has reference points to all the data stored in various parts in
the network.
[0516] There are questions that require the AI program to extract
information from long-term memory. The example below illustrates
this point.
[0517] The example in FIG. 80A illustrates how long it took the
robot to complete a task. The robot first searches for the movie
frames regarding the building of a ship. Then it extracted the time
it took from start to finish and use that information to answer
the question. As always, this example will be compared to similar
examples already stored in memory and the AI will determine wither
or not there are patterns involved. FIG. 80B is the universal pathway
used to answer these type of questions.
[0518] Everything in the pattern can be in a fuzzy range. For example,
the question "how long did it take you to finish M1" can
be represented as "how long did it take you to accomplish M1"
or "you worked on M1 for how long?". So, everything in
the sentences can be in a fuzzy range and doesn't have to be exactly
as the pattern. Everything from the sentences, to the image layers,
to sound, and even the position of the image layers can be in terms
of fuzzy logic.
[0519] Let's combine both internal function C and internal function
D and give an example of both functions working together to answer
a question. As mentioned above, all internal functions can be combined
together to look for information. The pattern can be simple with
one internal function or it can be complex with multiple internal
functions working together to find information.
[0520] FIGS. 81A-81B are diagrams showing two internal functions:
finding data from the 3-d environment and rewinding and fast-forwarding
in long term memory to get information. The example in FIGS. 81A-81B
uses both the long-term memory and the 3-d environment to look for
information. First the AI program looks for the movie frames concerning
Jessica's mouse. Then it extracts the movie frames from the long-term
memory. Next, it extracts the information that it needs from the
movie sequences (in this case it wants to know where Jessica's mouse
was put last). Finally, it takes this knowledge and answers the
question.
[0521] When many similar examples are trained the AI program will
understand the question in a fuzzy logic way. The universal pathway
will be created in terms of this question and answer situation (FIG.
82A-82B).
[0522] All internal functions are assigned to its appropriate places
in this universal pathway (FIGS. 82A-82B). Before answering the
question the AI will use internal function D (searching for information
in long-term memory). Then it takes particular movie sequences and
extract information from these frames using internal function C
(search for information in the 3-d environment). The assigning of
these internal functions to a particular moment in a pathway is
done by averaging similar pathways and finding the patterns. It's
kind of like reverse engineering what an event is and assuming what
internal functions were used to get a particular information. The
patterns are found and the event is assigned certain internal functions
to instruct the AI to find information in memory. This is how the
robot will be able to predict the future or find meaning to language.
And these things are all done through fuzzy logic.
[0523] E. Determining the Strength and the Weakness of Data in
Memory
[0524] In this example I will combine the assignment statement
and the strength and weakness of data in memory. In FIG. 75, the
"this is a mouse" example will be revisited. In order
to assign one object to another object the AI program has to encounter
these two objects many times before they can be assigned to each
other. For instance, if I wanted the robot to assign the sound "cat"
to the visual image of a cat, I would have to train the robot with
both objects repeatedly. Maybe after 20 sets of training the AI
program will understand that the sound "cat" is equal
to the visual image of cat. If you think about all the words in
the English language and how long it would take the AI program to
learn these words, it would be very overwhelming.
[0525] There is an alternative to this brute-force way of learning
words. The English language can be used to encapsulate patterns
and these patterns can be used to accomplish certain tasks that
would otherwise take a long time to finish. In FIG. 75, the "this
is a mouse" example is designed to assign a word to a particular
image. Along with determining that two objects are equal it can
copy the connection strength of the two objects involved. This will
allow the AI program to encounter two objects once or twice and
the AI program "gets it". The robot understand that this
particular word identifies this particular image in the current
pathway. Instead of using the old method of training the word with
the image we have used sentences to represent assigning equality
among two objects. In other words instead of training the robot
20 times with the two objects we can use the sentence 2 to 3 times
before the robot understand a meaning to a word. The strength of
the two objects (word and image) are given the average strength
of all similar examples.
[0526] In FIG. 83 the sentence "this is a S1" is assigning
the word S1 (a variable) with the image layer in the frame. The
sentence will also assign the average strength of the connection
between the target object and the element object. In this case the
average weight of the connection is 90 pts. When the AI encounters
the sentence "this is a bat" and in the frame is an image
of a bat, the AI program (if it never saw a bat before) will create
the word "bat" in memory and it will store the bat image
close to the word "bat" with the connection weight set
at 90 points.
[0527] Conscious Thoughts and its Development
[0528] Up to this point the AI program can understand the meaning
to words/sentences and it can also create patterns in pathways that
can predict the future. The understanding of meaning to language
is also accompanied by fuzzy logic so that the meaning is more important
than the words/sentences that represent that meaning.
[0529] The material covered up to this point is important to the
understanding of conscious thoughts and how it is developed. The
conscious serves many purposes for the robot. It provides the robot
with valuable information about the environment, it gives meaning
to language, it tells facts about an object, it guides the robot
to solve arbitrary problems, it answers questions, and even provides
a conversation when the robot is bored. (some conscious thoughts
has very little to do with the 5 senses from the environment. This
will be explained further in later sections)
[0530] The idea behind the conscious is quite simple. The AI program
recognizes target objects from the current pathway and all the elements
from all target objects compete with one another to be activated
in memory. These activated element objects are the conscious thoughts
of the robot (FIG. 84).
[0531] FIGS. 84-85 are diagrams depicting target objects and activated
element objects. The arrows at the top of the timeline are target
objects and the arrows located on the bottom of the timeline are
activated element objects (FIG. 85). All the target objects and
all the activated element objects will have their element objects
extracted from memory and the rules program will decide which of
these associated element objects will be activated. Although activated
element objects don't have the same strength as target objects,
the rules program will consider the activated element objects too
(activated element objects have 1/4 the strength of target objects).
In FIG. 85 all target objects and activated element objects closest
to the current state will be considered first, while objects farther
away will be considered last. This also means that the objects closest
to the current state have higher consideration than objects farther
away from the current state. When I say objects I'm referring to
both target objects and activated element objects.
[0532] Fuzzy Logic is Very Important to the Rules Program
[0533] All data in memory is represented in terms of fuzzy logic.
Visual images or 360 degree sequential images of a visual object
has a fuzzy range of itself. A 360 degree floater of a cat will
identify all the cats in the world despite their physical appearances
such as size, color, weight, and age. The meaning to certain words/sentences
also has a fuzzy range of itself. People can say totally different
sentences but the sentences mean the same things (or roughly the
same things). You can even use sentences from two different languages
but the meaning of these sentences mean the same things.
[0534] The fuzzy logic is what brings order to chaos in the world
we live in. It is also a very powerful tool used by the rules program
to create conscious thoughts. The life we live in has infinite possibilities
and chances of encountering the same sequence of events twice is
impossible. However, we can encounter events in life in a similar
manner.
[0535] Because all the data is stored in terms of fuzzy logic and
each data has a hierarchical order of itself, the strength of pathways
depend on which pathways in the hierarchical order is strongest
and not the exact data itself. (Remember I said that the AI program
may not pick a 90 percent match compared to a 20 percent match).
The reason is because the strength of the pathway also matters in
the decision making process.
[0536] FIG. 86A is a block diagram showing sequential sentence
association. Data 230, data 232, and data 234 are the training examples.
[0537] If we train the 3 examples in FIG. 86A over and over again,
the AI program will have a strong connection between the first sentence
and the second sentence. Although the first sentence isn't the same
every time the meaning behind it is the same. The sequence is trained
in terms of fuzzy logic and the meaning (a hidden object) is more
important than what is actually sensed (the target object or the
sentence).
[0538] If we apply this example to the rules program the reader
will get an idea how the conscious works (FIG. 86B).
[0539] The sentence that was encountered by the AI program, "you
bought a blue key at the supermarket" (FIG. 86B) isn't a sentence
that was trained in memory. The three examples that were trained
were different sentences but they share the same meaning (meaning8).
[0540] Because there was a strong association between the two sequential
sentences in memory, when the AI encountered the target object "you
bought a blue key at the supermarket", the second sentence
was the second element object to activate. The meaning of the first
sentence was the first element object to activate. The reason for
this is because the meaning had stronger association to the target
object.
[0541] The second sentence can also be trained in a fuzzy logic
manner. Instead of an exact sentence, a meaning can be activated.
[0542] The TV Problem
[0543] The next example illustrates how human conscious is used
to create logic and reasoning. This example was taken out of a movie
that I was watching. An idea popped up in my head when I was watching
a scene where reasoning was needed to understand the situation.
I did some reverse-engineering on how the logic was created and
found out how reasoning happens in human beings. The diagram in
FIGS. 87A-87D demonstrate this form of logic.
[0544] In FIG. 87A the reasoning behind this situation is that
Jane told Dave not to watch TV on that day. When Jane came home
from work Dave said that he went to fix the antennae. The logic
behind T3 is that Jane knows that the antennae is attached to the
TV and that the TV must have been broken. The only way that the
TV broke is if Dave was watching TV and something happened to it.
The way that the AI program outputs the logic in T3 is by the lesson
I thought earlier about sentence association. The more times the
robot learns knowledge about a situation the more likely that knowledge
will be activated by the rules program. Knowledge could be any data
in memory, most notable sentences or movie sequences that include
sentences and words that have references to the movie sequence.
[0545] The knowledge from logic T3 is presented in FIG. 87B-87C.
These strong sentences were activated by the rules program and it
gave Jane the knowledge to come up with T3's logic.
[0546] The knowledge base 246 and 260 are lessons learned by teachers
or by observation (FIG. 87B-87C). They are just a bunch of sentences
and movie sequences that teach a person knowledge about a situation.
The objects within these knowledge base are strong so when one object
(first sentence) is recognized by the AI program the other object
(second sentence) in the situation will activate. In the example
in FIG. 87D, the situation is set up when Jane told Dave not to
watch TV on that day. Then 5 hours later Jane got off work and went
home. When she got home Dave told her that he went to fix the antennae.
The response she gave Dave comes from the logic above. That logic
is: The meaning of the sentence: I went to fix the antennae activated.
This meaning had strong association to knowledge base 246A which
activated the first sentence. Next, Jane activated the strongest
association to the first sentence which is the second sentence:
Dave (D1) was watching TV and the TV broke. Then Jane activated
knowledge base 260A where a previous event 268 triggers knowledge
base 260A. The decision to activate knowledge base 260A comes from
a pattern to extract knowledge from the past, in this case extract
the event where Jane told dave not to watch TV. The result is the
conscious thought: I told Dave (Z1) not to watch TV today (Z2).
The association that is attached to that is to say to Dave: "I
thought I said no TV today".
[0547] This example demonstrates reasoning in robots and how the
conscious is used to create this reasoning. Although this is a relatively
simple example, if you think about all the steps that are described
above and combine that with fuzzy logic then you will understand
how affective this form of reasoning is. The knowledge base can
be represented through fuzzy logic, the steps of recognizing the
objects from our environment can be done in fuzzy logic, and activating
element objects can be done in fuzzy logic.
[0548] The knowledge base of the program can be as long as you
want it too be. You can read an entire science book and all of that
knowledge will group itself based on their strongest association.
When that knowledge is recognized by the AI program the strongest
knowledge attached to it will activate.
[0549] How the AI Program Builds this Knowledge Base
[0550] The AI program learns knowledge by reading books. However,
before it can read a book it needs to understand all the words/sentences
and the grammar structure of a language. That is why it is so important
that the AI program have the ability to understand most of the words/sentences
from a language. Just like humans these robots have to learn knowledge
from a young age and slowly build all the neural pathways in memory.
[0551] Things like creativity are actually just lessons learned
in life. If the robot is drawing a picture all the strongest lessons
learned to draw a picture activates in the robot's mind and these
lessons instruct the robot to draw the picture. Although there are
some lessons that are created by the robot the majority of the lessons
are guided by teachers. A question or a statement is thought to
the robot and that question or statement is asking the robot what
its' preferences are. Questions or statements such as: "what
is your favorite color?", or "if the picture doesn't look
good use the eraser and try again". These aren't instructions
that the teacher gave to the robot to draw the picture, these are
statements or questions that ask the robot what it wants to do.
The answering of the questions or statements are the instructions
that is used to draw the picture.
[0552] Ideas and imagination is also part of the conscious. Just
like before, conscious thoughts that create ideas come from lessons
learned in the past by teachers. Ever since we were in grade school
the lessons learned by teachers guided us in terms of creativity.
Wither its coming up with a good essay or making a business plan
or drawing a picture, that creative side of a human being comes
from the average lessons thought by teachers. Statements like "we
need to come up with a new idea that we never thought of before",
is a very powerful statement because in order to answer this statement
you have to understand certain information. One of those information
is what kind of ideas have you explored in the past and what kinds
of ideas have you come up with in the past but didn't use. These
information are needed in order to come up with a response to the
statement. Creativity is a very complex form of intelligence and
in order to form a creative mind many years of learning must be
had. Creativity is also something that is encapsulated with many
forms of intelligent thoughts. The complexity is managed by sentences
and meaning of sentences.
[0553] I want to reemphasize one more thing because I think it
is very important to the understanding of how fuzzy logic works.
It isn't just words/sentences that are represented in a "fuzzy
logic" way, but entire situations where visual movie sequences
are accompanied by sentences to accomplish a goal. The knowledge
base doesn't just come from reading a book with text but reading
text along with pictures and diagrams and examples. Math books have
a lot of these examples and diagrams to solve a problem or a science
book have instructions in terms of pictures and text to point out
how experiments are carried out. The knowledge base will include
not only text but also visual movies that contain text to describe
things.
[0554] Expert in Writing Essays and Giving Speeches
[0555] The more you read and the more you understand how the grammar
works the easier it is to recognize and store the words/sentences.
The easier it is to recognize and store the words/sentences the
better the logic and reasoning for that intelligent being.
[0556] For something like writing an essay, it requires many tasks
working together in order to accomplish. First the understanding
of most words/sentences has to be understood. Then understanding
grammar and how words/sentences are structured in terms of language
rules. Next, you have to know how to write an essay--what are the
steps and rules in writing an essay such as identifying the topic,
how long does the essay have to be, what font size to use, what
are the paragraph indentations, the margins of the page, structure
of the paragraphs, and proofreading the text. Finally, there is
the imagination part of the essay. The writer has to come up with
ideas to write the essay. These ideas come from personal knowledge.
Just like how the robot can learn how to draw a picture it can learn
how to write an essay.
[0557] Giving speeches is also another task that is very complicated.
The speaker has to prepare the speech. The speaker also has to know
what is contained in the speech and how to give the speech. What
it boils down to is many many years of learning the English language
and learning how to give a speech before such a task can be accomplished.
As the robot learns more the knowledge in memory builds on itself
and the complexity of any problem is managed by encapsulation.
[0558] Conscious Thoughts Part 2
[0559] In the previous sections we discuss how text (sound or visual
text) can be used to create reasoning. In this section instead of
using only words/sentences I have decided to demonstrate intelligence
using words/sentences and visual movie sequences. A math problem
is something that can't be solved through text alone. It can only
be solved through visual movies and words/sentences.
[0560] FIGS. 88A-88B are diagrams showing an example of an addition
problem. Most of the sentences are learned previously such as sentences
like "take the answer, 8, and put it under the column".
The sentence instructs the robot to identify the number 8 and then
copy that 8 under the column. This sentence was learned previously
and the understanding of the sentence means the robot can carry
out the instructions. Another previously learned sentence is "take
the 1". This sentence is trying to focus the robot's eyes on
the number "1" on the visual environment. That number
"1" that was said in the sentence represents the visual
number "1" on the math problem. Other variations of the
sentence like "look at the number next to 1", means identification
of numbers in relation to the visual environment. These sentences
instruct the robot to focus and assign words in the sentence to
images in our environment. The meaning to these sentences uses hidden
objects and patterns (previously discussed).
[0561] Next, the AI program has to have many similar training examples
in memory so that the AI can find patterns and similarities between
all the training examples. The common traits within the hierarchical
pathways (called a floater) will be developed where all the data
are centered at the strongest hierarchical pathway creating a fuzzy
range of itself. Anything that falls within this fuzzy range will
be considered the same object.
[0562] FIGS. 89A-89B is a similar example to the math equation
above. The numbers are different and the sentences used to solve
the problem are different. These are different sentences but mean
the same things. The overall way of solving a similar math problem
is the same. It's just that certain variables are different and
the AI has to identify what is similar or same among all the training
examples.
[0563] This similar pathway to solve a multiplication problem is
a variance of the first example (FIGS. 88A-88B). The sentences used
are different (same meaning), the numbers used are different, the
timing of the sequences are different, and the way the numbers are
represented are different. All of these things will be averaged
out by the AI program and a universal pathway will be created in
order to solve this problem.
[0564] The timing of the problem is one factor to consider. The
AI will average out the time it took to solve this problem (FIG.
90).
[0565] When the average is created the AI will void any discrepancies
in terms of time. However, the time it takes for a math problem
should fall within the average time in the floater to be considered
a pathway in this floater.
[0566] Another lesson that I want to note is that the hierarchical
order of image layers must be considered (FIG. 91). If a number
2 was identified and a number 4 is identified that means that the
most common learned group is the word: "number". Both
numbers is considered the same at the learned group "number".
This is important because when the AI averages out the pathways
the image layers contained in the pathway are purely numbers. They
aren't alphabets, or toothpicks, or pencils. The elements in a multiplication
problem are purely numbers.
[0567] The self-organization part of the program will average out
the image layers in similar pathways and creating a universal pathway.
FIG. 93 is one example. N in this case stands for number (block
270). Sentences will also be averaged out where the hierarchical
meaning of the sentence is established and not the sentences that
represent that meaning.
[0568] Using Visual Movies and Words/Sentences to Learn Other Knowledge
[0569] There is another way of learning learned groups besides
the material I have covered. Learned groups are language that classify
things around us. We associate a 5 sense object with a certain word.
The word can be anything in our environment. That 5 sense object
doesn't even have to be similar in physical appearance. The word
animal encases many visual objects in our environment. These objects
aren't even remotely similar to each other in physical appearance.
A dog and a rat doesn't look similar or a cow and a giraffe doesn't
look similar. Despite physical appearances all these visual objects
are classified as animals (a learned group).
[0570] The previous way of assigning a visual object with a learned
group is by having the AI find association between two objects (FIG.
92). If the two objects fall within the same assign threshold then
both objects are considered identical. Usually, words are used as
learned groups to classify visual objects.
[0571] The second way of learning learned groups is by using visual
images and sentences to explain what a word means (In fact, there
are combinations of ways in which words can be assigned to a visual
object). Maybe by using a diagram to create associations between
word and visual objects learned groups can be created. FIGS. 94-95
are examples of several learned groups that can be represented by
visual diagrams and text. Images 272 and 278 are assigned to their
respective word/s 274 and 276
[0572] Not only can learned groups be represented from movie sequences
but also a hierarchical tree. A hierarchical tree of mammals can
be created and understood by the viewer (FIGS. 96-97). A hierarchical
tree of a family can be understood by the viewer. Within the tree
sentences can be used to explain what functions each element in
the tree serves and how it relates to other elements in the hierarchical
tree. Words/sentences alone can't explain what hierarchical trees
are. But these diagrams can give the viewer the understanding of
a hierarchical tree from a learned perspective.
[0573] Referring to FIG. 96, when questions are asked such as:
"Are humans and animals mammals?", the patterns involved
to answer the question (future prediction) comes from this diagram.
Facts about the diagram pops up and the robot uses these facts to
answer the question. Patterns are found between similar examples
and the instructions to answer the question will be in the patterns.
Facts like "humans and animals come from the same group, mammals"
can be used to answer the question. If the visual diagram above
was on a textbook and one of the assignments given by the teacher
is to answer this question: do animals and humans have a female
type and a male type?, the answering of the question will require
the robot to observe the diagram and read the text. Based on what
it learned it can use the knowledge to answer the question "do
animals and humans have a female type and a male type". Such
behavior to answer this question require many training examples.
As usual the complexity is managed by the AI program.
[0574] On the other hand, let's use a family tree as another example
to demonstrate intelligence. Imagine that the diagram in FIG. 98
was presented to you by a teacher. And this teacher gave you facts
about all the elements in the family tree such as: "the father
is always male", "the mother is always female", "son's
and daughters belong to the father and mother", "the mother
and father are both parents to son's and daughters".
[0575] Based on all these facts about the diagram and repeatedly
teaching people what the relationship between the elements in the
diagram are, the robot is able to learn what a family tree is. Answering
of questions related to this family tree can happen by using this
diagram from memory and using the facts that are activated by this
diagram.
[0576] Common Sense Knowledge or Observation Consciousness
[0577] Learning to observe the way people behave and act is very
valuable to intelligence. Also, observing a situation and what the
appropriate actions are is very valuable to intelligence. Common
sense knowledge is what most AI scientists call this field of research.
The ability for machines to understand knowledge that humans have
is quite complicated. When someone drops food on the ground, a human
knows that the food is contaminated and can't be eaten. When it
rains a human will take shelter, when humans smell smoke he/she
will run out of the house. These are common sense knowledge that
humans have. This type of knowledge was learned from the day you
were born to your current state. Common sense knowledge is actually
the ability to learn to observe a situation and to have a teacher
teach you what that situation is.
[0578] The best example of observation is from my English class.
I was studying Shakespeare for that semester and I had to read Hamlet.
In the book there was a line that I didn't understand and I wanted
clarity by asking the professor. The sentence I was confused with
was: "more matter and less art". From a human point of
view this sentence makes no sense. But after asking the professor
what it means and using a form of complex logic I figured it out.
The statement: "more matter and less art" means "get
to the point".
[0579] I use this example because in Shakespeare's plays, the language
he uses is different from the language we use today. In order to
learn the language I had to analyze the sentences, word for word,
and have a teacher tell me what different words mean and how that
word relates to other words in the sentence. I understood the complex
sentence from observing an explanation of the meaning.
[0580] The next time I read the sentence: "more matter and
less art" the meaning "get to the point" pops up
in my mind.
[0581] Spilling Milk Example
[0582] Imagine there was a scene where a boy from Korea who is
holding a bottle of milk. The boy comes from a very poor family.
It took the boy two hours to get to the market place to buy the
milk. The boy runs and trips spilling the milk all over the floor.
The boy gets up and looks at the empty bottle of milk, and then
the boy begins to cry. Based on this scene an intelligent person
would understand that the boy did not cry because he tripped and
fell to the ground. I'm sure it was a little painful but the boy
didn't cry because of the fall. The boy cried because the milk was
spilled all over the floor and the milk was gone. Since the boy
is poor he and his family won't have any food for the rest of the
day.
[0583] As students in school we learned how to observe a situation
and either hear what people think about the situation or we can
voice our own opinion about the situation. This is where the conscious
of intelligent thought is produced. The collective voices of not
only the teacher but the other students who critiqued about the
situation is stored in memory in a fuzzy logic way (FIG. 99A). All
the sentences said during the situation are averaged out and what
remain are the strongest average sentences for that given situation.
[0584] Based on the conversation about the situation the robot
will store all these sentences in memory and average the data. The
diagram in FIG. 99B will show a similar example to the boy spilling
the milk and the teacher's and students' responses to the same situation.
All the responses are stored in terms of fuzzy logic
[0585] In the second example (FIG. 99B) the speakers are different
and the way that the speaker says the sentences are different. The
timing of the sentences are said at different times too. The important
thing is that the meaning is the same. And because the meaning is
the same the computer can average all similar examples and come
up with a universal pathway. What will activate is the meaning of
something instead of the exact sentence that was encountered. This
is the essence of fuzzy logic.
[0586] Observing and listening to people's opinions about a situation
is very important to common sense knowledge. The brain will have
to know facts about a situation and know what to say and do next.
The material learned in school is vital to the way conscious thoughts
are activated. If someone walks into a classroom with a black eye,
people will critique, assume, and guess using logical analyzes of
the person. They can assume this person got into a fight yesterday
and got punched on the eye. It could possibly be that it was an
accident. Whatever the circumstance is by analyzing this person
and his behavior human beings can assume what happened.
[0587] Another example is if someone is sick. We learned that if
someone is sick we have to take measures to make sure that there
is no contact between the sick person and us. The reason for this
is because sickness can be spread among humans. We learned how diseases
are spread and the flu is spread through contact with the sick person.
Because we discussed the situation when a person is sick and how
to respond to this situation we know what to do or we know how to
think consciously when such an event occurs. In some sense this
form of analyzing a situation can be used to predict the future.
In previous lessons I taught about how the AI program follows the
strongest future pathway in order to predict the future. This is
the second way in which the AI program can predict the future--by
using sentences and logical analyzes of the current situation.
[0588] We learned how to respond to danger when it occurs because
the teachers thought us how to respond. For example, we know that
alligators are dangerous. We didn't learn this lesson by having
an alligator bite us, we learned it by lessons thought to us in
school. Sometimes pain and pleasure decides things but this time
it's using sentences and logical analyzes to tell us what to do
in the future. "when we see an alligator or any dangerous reptile
what should we do?" "we should run away and get help".
That conscious thought instructs the robot in what kind of action
to take in the future.
[0589] Observation by Watching TV
[0590] Another form of intelligence is observing a situation by
watching TV shows. Copying what to say and when to say it as the
show is interpreted by the robot. Observing how others interpret
a situation and either agreeing with them or disagreeing with them.
Many logical thinking is done by watching TV because the scripts
are well planned out. In fact, most of the learning we get comes
from watching TV and copying the things that happened during the
show.
[0591] By watching the movie and making personal opinions about
a situation we are learning to analyze a situation. During the movie
there might also be people critiquing on the situation so you can
get their opinion on the situation.
[0592] Copying the way actors behave, say things, and act are another
factor that can be considered when watching TV. We tend to emulate
certain people that we look up to. Some might be people from real
life, but others are actors and actresses on TV. We take the lines
that we find dynamic and we copy them. If we like the way certain
actors/actresses dress then we copy them. If we find their line
of work interesting then we try to work in their field. So, lots
of behavior and decision making are done by watching and emulating
what we see on TV.
[0593] Markers and How Sentences Play a Role in Identifying Pathways
[0594] Sentences are just markers on the pathway and are not considered
an entire pathway. Sentences don't actually encapsulate entire situations
(movie sequences). It simply gives the AI program a marker in a
particular unique area in memory and the unique area happens to
be the only pathway that contains the sentence. I will be using
the ABC block example again (FIG. 41A). At the beginning of the
problem is a sentence that identifies that pathway: "I want
you to stack the blocks up starting with C then B and finally A".
[0595] This sentence serves as the marker to identify the entire
pathway. In fact, this sentence is so unique that only this pathway
contain the sentence and no other. In previous lessons I stated
that a sentence has to be identified according to the situation.
This doesn't mean that the sentence identifies the entire pathway
as the meaning. It just means that at that moment a hidden object
(meaning1) is activated and this meaning1 is just a pattern that
tells the robot what to expect in the future as a result of the
sentence.
[0596] If you look in FIG. 41A every sentence in the problem is
a marker. Every sentence is unique only to a certain pathway. By
identifying the sentence, the pathway is also identified. Each marker
might belong to other pathways but it's the combined sequence of
the markers that make the pathway unique.
[0597] Each letter in FIG. 100 represents a sentence (marker).
If you wanted to match the pathway in FIG. 101 to one of the pathways
in memory (FIG. 100) then the AI has to find the best match. According
to the match all three pathways have letter "A", therefore
the only way to choose a pathway is to look at their powerpts. Since
pathway 1 has the highest points (96 pts) then that is the pathway
the computer will choose (AZX).
[0598] However, if the pathway is like the example in FIG. 102
then the AI program will pick the best sequential match that contains
sentence "A" and sentence "B". The more sequences
the AI is allowed to search the more accurate the match will become.
In this case the pathway is so unique to ABC that there is only
one pathway it belongs to (the 2.sup.nd pathway).
[0599] Back to Knowledge Base and How it Works
[0600] The diagrams in FIG. 103A are sequential events that happen.
Imagine that each letter is a word in a sentence and that the robot
is reading in text from a book. Notice that the grey blocks: ABC
block and CKNW are outlined. I wanted the readers to be aware of
these two blocks.
[0601] In the next diagram (FIG. 103B) the machine recognized ABC
and the current state is at: CKN. Based on CKN the rules program
activated CKNW. The stereotype CKNW is attached to object CKN and
that is why it was activated.
[0602] Target object ABC and target object CKN are trying to compete
with one another to activate their respective element objects (FIGS.
103C-103D). They both share CKNW as an element object. This makes
the element object CKNW stronger. The rules program activated CKNW
as the element object at that moment because that was the strongest
element object based on the current data.
[0603] Notice that in the knowledge base that ABC and CKNW are
not even trained sequentially. They are far away from each other.
But all three sequential training example has object ABC first then
object CKNW. Because both objects are trained together each object
is associated with the other object.
[0604] Decision Making and Planning Tasks
[0605] There are many different levels on decision making and each
level influences the way the robot makes decisions. Below is an
outline of the level of factors that will influence the AI program
in terms of decision making.
[0606] Levels of Decision Making: [0607] 1. innate reflexes based
on pain--when a person is in great pain reflexes are most likely
to trigger. These reflexes are wired into pain so that when the
pain reaches a certain point it triggers the reflex. No conscious
decision was needed to trigger this action. Some of these innate
reflexes are: when a person is in great pain he/she yells out loud,
when the knee cap is hit with a hammer the leg moves automatically,
etc. [0608] 2. Learned decisions based on past knowledge--the conscious
guides the robot to make decisions. These decisions are either based
on future predictions or logical decision making. [0609] 3. Pain
and pleasure built into the robot--attractiveness or ugliness, physical
pain (degree of pain) and physical pleasure (degree of pleasure)
are factors that the robot uses to make decisions. Is the robot
going to eat lobster for dinner (the robot loves lobsters) or is
the robot going to eat rice? (the robot eats rice only if he has
to). These pain/pleasure factors that are built into the robot will
make decisions. [0610] 4. Daily routine--Learned things that the
robot was thought everyday by teachers are factors in decision making.
Some of these daily routines are so natural that the robot doesn't
need to make a decision to do them. Daily routines such as: waking
up in the morning, brushing your teeth, eating 3 meals a day based
on the time, going to sleep at night, using the bathroom when you
need to go, and going to work or school on weekdays.
[0611] These are the levels of decision making. The higher levels
overshadow the lower levels in terms of decision making. For example,
innate pain overshadows learned decisions because innate pain is
a reflex and is triggered by pain while learned decisions uses conscious
thoughts to make decision. In other words, innate pain is triggered
without conscious thought and overshadows learned decisions.
[0612] Another example is learned decisions can overshadow pain
and pleasure. This form of pain and pleasure doesn't trigger reflexes,
it's just a lower degree of pain/pleasure from innate reflexes--a
degree where the robot can manage the pain. If a person has an itch
on his butt and this person is walking on the street, the person
can make a decision to scratch his butt or not. He can wait until
he gets to a private area before scratching his butt. This is one
demonstration of learned decision having higher priority than pain/pleasure.
Even something like using the bathroom require learned decisions.
If you have to go the pain is unbearable. However, you can't take
a dump on the street or in a classroom. You have to make a decision
to go to the bathroom and take a dump. Even though the pain is so
great the learned decisions guided the robot to take the appropriate
actions.
[0613] Pain and pleasure is another factor that is used for hidden
objects. The AI finds these patterns and wire pathways with pain
and pleasure. The strongest pathways have their powerpts strengthened
because it's wired to pleasure and the weak pathways have their
powerpts lowered because it's wired to pain. The learned decision
encapsulates pain/pleasure to plan out tasks and make decisions
for the robot. The main function of decision making is always to
pursue pathways that lead to pleasure and stay away from pathways
that lead to pain.
[0614] Daily routines such as brushing your teeth, sleeping at
9 pm, and waking up at 7 am are just things that we learn everyday
and this type of learning is so normal that we do them without thinking.
Learned decision can overshadow these things because we can control
when we sleep by conscious thought. Instead of 9 pm we can sleep
at 2 am. Instead of eating cereal for breakfast we can eat a hamburger.
This daily routine is also another factor that can be encapsulated
into learned decisions to plan out tasks and make decisions for
the robot. The AI program finds patterns concerning daily routines
and use this pattern in a hidden object. This hidden object will
then be assigned to words/sentences as meaning of words/sentences.
[0615] Planning Out Tasks and Interruptions of Tasks
[0616] In modern day AI techniques, planning out tasks uses a combination
of language parsers, discrete mathematics, probability theories,
and recursions. My method of planning out tasks uses the conscious.
Everything from planning a task, decision making, task interruption
and probability of task is all managed by one thing: conscious thoughts.
[0617] In this section I will demonstrate how tasks are planned
out and how interruptions of tasks are solved. So far you have learned
that the conscious does many functions for the robot. Functions
such as provide meaning to words/sentences, give information about
objects, guide the AI program to solve arbitrary problems, and provide
information about a situation. Now, an addition to these functions,
the conscious can plan out tasks and solve interruptions of tasks.
[0618] There are two ways of accomplishing planning of tasks. The
two ways will be outlined and detailed demonstrations will be given.
[0619] 1.sup.st Way of Planning Out Tasks:
[0620] FIGS. 104-107 are diagrams showing the process of planning
tasks and managing interrupted tasks via language. Referring to
FIG. 104, pathways in memory are either continuous or non-continuous
(pathway 280 and 282). The continuous pathway is a pathway that
can be followed sequentially. The non-continuous pathway is a fabricated
pathway that jumps around in memory. As the AI follows a pathway
it will keep a note of wither a pathway is continuous or not. It
will also indicate where the pathway jump to or from.
[0621] So the AI program was following pathway one (P1), then it
jumped to pathway2 (P2), followed P2 for awhile then it jumped back
to pathway one (P1). This is one pattern that will be used to self-organize
similar pathways in memory. Imagine if P1 represent the ABC block
problem and P2 represent an interruption by a student. While the
robot is trying to solve the ABC block problem a student in the
classroom interrupted him to sign his name on a piece of paper.
After the robot finishes signing his name he goes back to the ABC
block problem and continues where he left off.
[0622] Conscious thoughts guide the student to go back to the task
it was previously doing. After the interrupted task is completed
teachers will teach the robot using sentences to go back to what
the robot was doing before--go back to the first task and continue
where it was before the interruption.
[0623] Before the robot decides to accomplish the interrupted task
teachers can teach it wither to do this interrupted task or not.
The teachers can teach the robot not to do the interrupted task
or to do the interrupted task. Maybe the teacher can set up some
kind of criteria of priority wither to abandon its current task
to accomplish another task.
[0624] FIG. 105A depict how conscious thoughts are used to plan
tasks and manage interrupted tasks. This sets up the rules for doing
another task and it also sets up the rules of what happens after
the interrupted task is completed.
[0625] The sentences, if understood by the robot, will carry out
the instructions to manage tasks. It will provide the robot with
rules to either do a second task or not. It will also provide the
robot with rules after the second task is done to either go back
to its previous task or continue on with another brand new task.
[0626] A universal pathway to manage tasks must be developed and
the only way to do this is by averaging similar pathways. The pattern
I stated at the beginning of this section is what will link similar
pathways together. The pattern is: The AI program was following
one pathway, then it jumped to another pathway, next, it jumped
back to the previous pathway. In addition to that the fuzzy logic
sentences that accompany a jump are included. FIG. 105B, FIG. 105C
and FIG. 105D are three examples of similar pathways and these pathways
are averaged out and a universal pathway is created.
[0627] In Ex. 1, Ex. 2, and Ex. 3 the situation is very similar
(FIG. 105B-105D). Although the tasks to be done are different the
way that the pathways jump around is the same. Also, the meaning
to the sentences in all three pathways are either similar or the
same. After self-organizing similar pathways in memory a universal
pathway is created. This universal pathway states that the AI program
was following pathway U1 and then it was instructed to jump to pathway
U2. After U2 is completed it was instructed to jump back to U1 and
continue where it left off (FIG. 105E).
[0628] U1 and U2 are pathway variables and the pathways can be
anything. U1 can be a math problem, or riding a bike, or taking
a test. U2 can be a conversion, or it can be using the bathroom,
or eating a piece of candy.
[0629] 2.sup.nd Way of Planning Tasks:
[0630] The second way of planning out tasks is virtually identical
to the first way. I simple add in the learned groups to identify
what the pathways are. Usually a sentence to identify the task is
crucial; other times it's just a combination of sentences to describe
the task that is crucial. The sentences don't represent the entire
pathway it's just a marker to identify a unique pathway (discussed
earlier). The AI program will use these unique markers (in the form
of sentences or meaning of sentences) as the identification of the
pathway. The sentences or its meaning will go through self-organization
just like all the other sentences and a universal pathway will result.
FIGS. 106A-106C are similar examples from the first way of planning
a task. I simply included sentences to identify what the pathways
are.
[0631] Referring to FIGS. 106A-106C, all the identification of
pathways in the form of sentences and meaning of sentences serve
as a marker on the pathway. The averaging of these markers will
also include any hierarchical order. For example, ex. 2 and ex.
3 are grouped in solving math problems. Although ex. 1 isn't remotely
close to ex. 2 and ex. 3, they are grouped together because the
overall pathways have similar events. Ex. 1 will still be included
in the universal pathway.
[0632] Within the universal pathway are hierarchical groups that
organize similar pathways (FIG. 106D). The more the AI program learns
the more organized these pathways are. The universal pathway is
the center of the floater because this is the pathway that is shared
among many examples. The specific pathways within the floater represent
the fuzzy range of the floater.
[0633] All the pathways are encapsulated in the universal pathway
(FIG. 107). Since Ex. 2 and Ex. 3 are so similar they are grouped
closer together. Ex. 1 is farther away. If the AI program encounters
a problem that is similar to Ex. 2 then it will go into the universal
pathway first, then it will go into the math problem group and finally
go into Ex. 2.
[0634] On the other hand if the AI program encounters a pathway
similar to Ex. 1 (the ABC block) then it will go into the universal
pathway first, then it will go into the Ex. 1 group.
[0635] Other Topics
[0636] The next couple of paragraphs are just lessons that were
taken out of other sections in this patent because they were too
long. I have included them here because these are important lessons
that should be noted.
[0637] Learning to Delineate Image Layers from Pictures and Movie
Sequences
[0638] There is another way besides using an image processor to
cut out image layers from pictures and movie sequences. Finding
patterns between the image processor and how it dissects out image
layers and assigning this pattern to language is another way. The
machine can use language as a tool to cut out image layers from
pictures and movie sequences. If we preprogram all the various ways
in which an image processor can dissect out images from pictures
it would be impossible to program. But if we teach the image processor
what to cut out in the form of sentences and visual movies then
it will know how to cut out images from pictures and movie sequences.
[0639] When I say cut out image layers from pictures and movies
I don't mean just cutting out moving objects. What I mean is there
should be rules that the image process must follow to cut out certain
images. If I said, cut out the image in the picture with the dotted
lines, then the robot should cut out the image by following the
dotted line and cutting it out carefully. If I said cut out animals
from the picture then the robot should cut out all the animals from
the picture. If I said cut out the images next to the mail box then
the robot should identify the mail box and then cut out the image
that is next to it.
[0640] By using language and visual representation we can guide
the robot to delineate any image from a picture or a movie sequence.
All the rules are communicated through language and the understanding
of language allows the robot to carry out the instructions. This
is the ultimate type of image processor that anyone can ever build.
[0641] Also, intelligence has lots to do with what rules the image
processor needs in order to carry out its instructions. The ability
to understand that a cat image is a called a "cat" and
a dog image is called a "dog". Being able to identify
situations like the dog jumped over the cat, is vital to the Image
processor. What if there was an instruction given to the robot that
said: "cut out the image that the dog jumped over". Since
the cat is the image that the dog jumped over then the cat is the
image the robot has to cut out.
[0642] We can use a finger to point to a particular image in a
picture. We can also use a laser to point to a particular image
in a picture. We can outline the image in the picture using the
laser. Or we can use an outliner like a digital outliner to delineate
an image from a picture. Learning sentences and movie sequences
and associating these things with a particular way of delineating
images from pictures and movies is one form of intelligent image
processing.
[0643] In fact, we can tell the robot what to do with the image
when it identifies this image in a picture or movie sequence. For
a real life picture we can have the robot cut out the image. If
it's an image in a picture on a computer monitor we can tell the
robot to erase the image using a mouse. If it's an image on a chalkboard
we can tell it to physically erase that image. If the image is on
a piece of paper we can tell the robot to white out the image. If
the image is on a coloring book we can tell the robot to color the
image. So the way that the robot treats the image in the picture
or movie sequence is arbitrary. Using language and all the mechanics
of the robot's body the different ways that the image can be handled
will be up to the programmer's imagination.
[0644] Reading a Book
[0645] When the AI program is reading a book he is actually fabricating
a movie sequence in his mind based on what he is reading (FIG. 108).
Every word that the robot reads in will activate a sequence of movies
that will tell the robot what is happening in the story. By the
time the robot finishes reading the story he will have an understanding
of the story not in terms of words/sentences but through a movie
that was fabricated based on the text in the book.
[0646] This fabricated movie will consist of snapshots of pictures
and movie sequences that are activated by the meaning of the words/sentences.
Although the fabricated movie won't be like a streaming DVD quality
movie the snapshots give an idea of what is happening in the story.
These fabricated movie sequences will be used to recall information
about the things the robot read. Questions that are asked about
the story depend on this fabricated movie in order to answer.
[0647] The foregoing has outlined, in general, the physical aspects
of the invention and is to serve as an aid to better understanding
the intended use and application of the invention. In reference
to such, there is to be a clear understanding that the present invention
is not limited to the method or detail of construction, fabrication,
material, or application of use described and illustrated herein.
Any other variation of fabrication, use, or application should be
considered apparent as an alternative embodiment of the present
invention.
|