Intelligent pathways
Learning from
childhood to adulthood and how the robot’s pathways
become more complex
The robot learns
information in terms of a bootstrapping process.
What this means is that it has to learn information
through stages in life. First, the robot has to
learn things in terms of a baby because there are no
knowledge predefined in the robot’s brain. Next, we
send it off to school, starting with elementary
school. Then, it has to acquire higher level
knowledge by attending high school. And finally, the
robot has to attend college, where its highest
knowledge is acquired. As it learns, information in the robot’s
brain builds on top of each other to form complex
intelligence.
Just like how we
humans have to learn to walk, to talk, to move, to
eat, these robots have to go through life the same
way. Let’s illustrate the gradual forming of simple
pathways into complex pathways by outlining a series
of stages. There are basically 7 stages of learning
that the robot has to master before having the
ability to do mundane human tasks. And these stages
are learned sequentially and they are not learned
individually.
1. innate reflexes
2. trained to do
things
3. sequential events
4. sentence commands
5. give robot option
commands
6. practice makes
perfect
7. copy other peoples
behavior
1. innate reflexes
In this stage the
robot will learn all the different objects that are
in the environment from the 5 senses: sight, sound,
taste, touch and smell. The robot was born into this
world without knowing that we live in a 3-d world.
It has to learn and understand what objects look like
in 3-d. Things like cat, dog, table, chair, red,
blue, car, house, I, her, him, loud, soft etc. are
learned and stored in memory.
Next, the robot will
start to move its arms and legs from innate built-in
reflexes. Movement of the arms, legs, mouth; and
controlling the vocal cords are things that the robot
must learn first (motor functions). These
experiences are 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 compared to 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. The meaning of objects will also be
established.
2. Trained to do
things
This part is where a
teacher will guide the robot to do things that are
appropriate and to force the robot to learn things
that a baby supposed to know (FIG. 28A). Things like
walking, and grabbing objects, 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 it wants to get from one
destination to another. Writing using a pencil is
important because the robot must learn to communicate
with people by written words. Things like walking
and writing and speaking must be learned by a guide
because we can’t pre-program the robot to learn these
things.
Although the guide
isn’t something we want to store in memory, the point
is that the more we guide the robot, the stronger the
desired created pathways will be. 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.
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.
3. Sequential events
In this stage the
robot begins to learn how objects interact with each
other. When two objects hit one another both objects
suffer, when the robot fall down it’s painful, when a
solid object is grabbed it has the same shape, but if
a soft object is grabbed the object has a bent 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 pathways 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 learned many times before
will be ignored because the robot 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’s real.
4. Sentence commands
This part will
require the robot to know basic grammar like the
names of popular objects found in the environment.
These basic grammar must be taught to the robot and
understood. The rules program will do the rest by
assigning the meaning (words) to a given object.
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 say,
“sit down”, the robot will sit. When you say: “pick
up the book”, the robot will pick up the book. When
you say: “read the first paragraph”, the robot will
read 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 might not understand a command and make
mistakes, but having a voice in the head that tells
the robot to do things hasn’t been created yet.
5. Giving the robot
option commands
This part is an
extension of the last stage. Instead of saying a
word and letting the robot do things, we can add
words/sentences intto the pathways and let the robot
decide what it wants to do. This is very affective
because using words to give commands will control the
pathway and to ultimately form state machines. In
this example, the pathway formed an if-then
statement.
So, this command
pathway represents 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 robot’s memory
there are powerpoints that determine an objects
worth. Pattern finding will tap into that and pick
the one with the highest 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.
6. Practice makes
perfect
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.
Referring to the
diagram below, 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 options can be added to this
practice pathway like, if you practiced 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 (a tree comprises one or more interconnected pathway).
So, an intelligent pathway can grow as it learns more and knowledge can grow like a tree, whereby weak branches fall off, new branches are formed, and existing branches can weaken or strengthen.
Another point worth mentioning is that language, in terms of words and sentences, can represent entire trees.
The words or sentences stored in pathways mark the beginning of a pathway (or trees). Because words/sentences can represent pathways, we can use words/sentences as reference pointers to pathways.
Let's say pathwayA contains this sentence, "pick up the paper" and this sentence references pathway4, instead of copying pathway4 into pathwayA, the robot's brain can use the sentence as a reference pointer so a copy of pathway4 is not stored into pathwayA.
This method minimizes repeated storage of pathways/data in memory.
Instead of following
commands given by external people, there are other
factors to consider before the robot takes action.
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 a 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 taught to the robot by
society.
7. Copy other
peoples’ behavior
This part is a very
powerful tool used to learn things. We can go ahead
and train a pathway that will allow the robot to copy
certain things from what it sees. 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, should do. 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 learned by
copying other people in our environment.
Adaptive intelligent
pathways
FIG. 34 depicts an adaptive pathway to beat a boss in the videogame, Contra.
The strategy that I use on all bosses and on all videogames is to first observe what the boss is capable of doing. Next, I identify any weak spots on the boss and determine the sequence of actions I have to take in order to hurt the boss.
Once a strategy is found and I identify the weakness of the boss, I repeat this strategy over and over again until the boss is defeated. I have been playing videogames for a long time, starting with Atari and onward,
I have never fought with a boss that didn't have a pattern. All bosses in all games I ever played had predictable patterns.
FIG. 34
According to FIG. 34, the player (which is the robot) is trying to find a weak spot on a boss. He tries one strategy, "hit the boss on the head" and that didn't work.
The strategies the player uses and didn't work are remembered so these strategies are not repeated. Next, the player brainstorms other ideas and decides to, "hit the boss on the back of the leg when he lands to the left", this strategy works.
This tells the player that one of the strategies to hurt the boss is to wait for the boss to land on the left of the screen and hit the boss on the leg. This adaptive pathway was designed to try strategies on a boss and to keep strategies that work and delete strategies that don't work.
After playing with the boss using 10-15 tries, the player has multiple strategies that it can use to hurt the boss.
The next step is to string these strategies together and determine the most optimal way, the fastest way, to defeat this boss.
The player keeps doing this until it discovers the pattern, the best strategies, to defeat the boss. The more it plays the easier it is to beat the boss. This adaptive pathway is universal and can be used to play any boss for any videogame.
Basically, what this adaptive pathway does is it generates a computer program inside the robot's conscious to find a strategy to beat a given boss.
It generates new strategies it can use to hurt the boss, it rememembers all failed strategies, it focuses on success strategies, it discovers the linear strategies to use in order to maximize the hurt for the boss,
and finally, it decides on the best linear strategies to destroy the boss. Furthermore, this computer program inside the robot's conscious can apply to any boss for any videogame console.
All intelligent pathways are structured hierarchically, and the pathways go from general to specific.
For example this adaptive pathway to beat a boss in a videogame, at the higher levels, can be used to design a poster. A graphic designer might select a pathway to try new ideas and to use them in a poster.
The good ideas that work will be used and the bad ideas will be discarded. Another example is doing a science experiment, this adaptive pathway can be used
to try a strategy until the experiment is a success. It can also be used to solve math problems, or to shop for the cheapest groceries.
Complex intelligent
pathways
Pathways from the
robot’s brain builds on itself as it learns more from
teachers. Even something as complex as representing
a hierarchy system can be generated by an intelligent
pathway. Things like parent-child relationships, who
is the grandfather of the family, or what does having
a brother really mean, can be represented by
intelligent pathways. When people say “that’s your
father”, there are lots of complex things we need to
know before we can understand what the sentence
really means. These complex things, include: “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.
The whole idea is to
use these intelligent pathways to form static or
linear state machines (computer programs). We need
the pathways to form: conditional loops, discrete
math statements, classes, methods, static variables,
recursions, various data structures, operator
functions, graph searches, etc.
Intelligent pathways
construct a computer program for a given situation;
it should have the ability to modify the computer
program as it learns more. All aspects of the
computer program formed by intelligent pathways can
be changed. These aspects include: variables,
frames/slots, classes, functions, conditional-loops,
recursions, search functions, data compatibility, and
so forth. For example, if the government changes the
laws for the game of basketball, the robot has to
learn the new laws and modify existing laws in memory
regarding basketball. In other words, the robot has
to modify intelligent pathways in memory that
represent linear procedures to play basketball.
Also, it has to modify knowledge stored in the brain
on basketball such as, the rules, facts, procedures,
and playing positions.
Driving a car is one
example of a human task. When the robot needs to
drive a car, its conscious will extract the necessary
knowledge in order to drive. The conscious will
construct one or more interconnected computer
programs to drive a car.
The intelligent
pathway to drive a car was created from all knowledge
learned about driving. Knowledge would include
lessons from teachers, books read on driving, trial
and error experiences driving a car, driving
examinations, and TV lectures on driving. The
totality of information learned on driving constructs
intelligent pathways in the robot’s brain to drive on
the road and highways.
The brain of the
robot is so well developed that it can learn to do
any human task via reading books or through trial and
error. If the robot wants to be a doctor he has to
go to medical school, if the robot wants to be a
computer scientists he has to go to a 4-year college,
if the robot wants to be a graphic designer, he has
to go to art school, etc. The knowledge learned on
each subject matter will be stored in memory and
intelligent pathways will be formed. Both knowledge
about the subject matter and the intelligent pathways
to extract and manipulate data are stored in the
robot’s brain.
Even the knowledge
learned through books or teacher lectures are
selectively extracted and stored in the brain in an
organize way. For example, if the robot reads a
controversial article, he will select only the facts
from the article that he believes is true and he will
try not to store false information.
There might be
ambiguous books and the author does a poor job in
explaining things. The robot can read vague
information and translate it into a form that is
understandable to the robot and store that
information in memory. Sometimes the robot can summarize a long book into short notes, so he doesn't have to store every detail of the book.
These summary notes can also organize details of the book by grouping information into manageable topics.
Storing minimal
amounts of data to do human tasks
The intelligent
pathways in the robot’s brain should be structured so
that only important information is stored and minor
or repeated information are eliminated. For example,
the for-loop in a pathway is used so that the robot’s
brain can jump back to a previous state. If the
for-loop wasn’t used, the intelligent pathway will
contain multiple copies of the same procedure. The
for-loop in pathways minimizes repeated data storage
to save disk space.
Just like creating
integrated circuits, the pathways in the robot’s
brain has to minimize repeated information. These
intelligent pathways have to be constructed in an
efficient manner.
In building computer
chips, a programmer has to determine what kind of
algorithm to use to minimize an integrated circuit.
In the case of intelligent pathways, pattern objects
contained in these intelligent pathways should
"self-minimize" intelligent pathways. The data in pathways
should organize itself so that data is void of
repeated copies, procedures are stored in optimal
areas, and data is stored in optimal areas so that
retrieval functions can access the data
efficiently.
As stated numerous
times before, the creation, modification, and
deletion of intelligent pathways in the robot’s brain
are done through lessons from school. The lessons by
teachers and books, teach the robot’s brain to form
and modify intelligent pathways.
Developing search
functions in intelligent pathways
In some ways, the
robot’s brain is no different from a human brain,
despite dissimilarities. Although I didn’t design
the robot’s brain with the same parts as a human
brain, the data in the robot’s brain self-organizes
to form parts that are similar to the human brain.
For example, the robot doesn’t have a predefined area
for language understanding. In a human brain, there
is one area that stores information on language
understanding and grammar. My robot’s brain wasn’t
predefined where the language part is
located, instead, as it learns grammar from school,
the language part develops somewhere in
the robot’s brain.
The same goes with
other parts of the human brain. I didn’t define
storage areas for each of the 5 senses into the
robot. However, as the robot learns more from the
environment, these 5 sense data are stored in their
respective senses. For example, the smell of an
orange is stored near the smell of an apple because
they have common traits, like smell data. So,
eventually, the robot will form different parts in
memory very similar to human brain parts. In the
robot brain, there will be a part for logic, language
understanding, visual cortex, sound, taste, touch,
smell, emotion, problem solving, and knowledge.
All information in
the robot’s brain is global. What this means is that
data in one part will intersect with data on other
parts. For example, the object apple might be
scattered throughout the entire brain. There might
me an apple stored in the visual cortex, the smell
area, the logic area or even the touch area. The
primary place to store an apple object is called
its masternode and the masternode has reference
pointers to other copies of the apple object located
in different areas of the brain. During the
self-organization phase, repeated data will be
eliminated or, at the least, minimized. Search
functions to locate the apple object, will first
locate the masternode. If the masternode is found,
its reference points will point to other copies of
the apple object, and each copy will be searched
based on priority.
The intelligent
pathways has to form search functions to search for
information in the robot‘s brain. These information
includes, searching for visual data, searching for
rules, searching for answers to questions, searching
for relevant information about one or more objects,
searching for ideas to solve a problem, searching for
memory experiences, searching for solutions to
problems, searching for steps to accomplish a task,
searching for future outcomes of a situation, and so
forth.
If you look at
google’s search engine, they spent 15 years
developing their algorithms. The robot’s brain
depend on intelligent pathways to search for
information for a given situation. These intelligent
pathways has to form search functions automatically,
without programmers inputting search algorithms.
Furthermore, the search functions in intelligent
pathways has to modify itself if the environment
changes.
If you think about
searching data in a massive database, this problem is
very complex. In database systems there might be
different topologies, media types, and data types.
My goal was to design the intelligent pathways so
that regardless of what data are stored in the
robot’s brain and how these data are organized, the
intelligent pathways will still be able to search for
information using the “best search methods“. Not
only does the intelligent pathways develop optimal
search methods, they will also develop the fastest
search methods.
<< notes