Following instructions using Human Level
Artificial Intelligence
Note: To make this website free to the public
please click on an ad to support my sponsors or you
can make a tax-deductable donation using Paypal
(click on the donation icon on the left).
This video shows a
robot reading and following technical instructions.
There are no sound in the video because I wanted to
show the viewers how the robot thinks while
reading/following instructions.
Before the robot can
read text, he has to understand how to read the text.
There is a specific way to read text, based on
several factors. These factors include: language,
media type, length of text and so forth. Reading a
Chinese book is totally different from reading an
English book. Reading a technical manual is different
from reading a comic book. There are rules and
procedures that the robot must follow in order to
successfully read text.
Next, the robot has
to read the text and create a fabricated movie, which
serves as the meaning to the text. This fabricated
movie consists of images, words, flow charts,
animation, diagrams to represent the meaning to the
words read.
In addition to the
meaning, the robot's conscious activates facts and
logic from the text read. It uses logic to identify
what tasks it has to do and to generate more facts on
the instructions. These facts generated are used to
truly understand the text read.
While the robot is
reading the instructions, knowledge pours into
primarily 4 containers: task container, rules
container, planning container, and identity
container. For example, if it identifies a task it
will insert the task into the task container. If it
identifies a rule it will insert the rule into the
rules container.
The process of
reading and following instructions is very complex.
Some of the instructions in the robot's pathways are
formed by repeated learning. Other instructions in
the robot's pathways are discovered by trial and
error. The knowledge that is poured into the
containers are based on patterns and these patterns
are discovered by repeated learning. No programmer is
needed to define what data to insert into the
containers.
Lastly, this form of
language understanding doesn't use: machine learning,
semantic networks, bayesian's network, planning
programs, expert programs, language parsers, grammar
rules, induction/deductions, predicate calculus,
decision trees, genetic programming, rule based
systems, and common sense systems, etc. Every
knowledge the robot has in his brain is learned from
teachers in school.
writing the software
program
part1: http://www.youtube.com/watch?v=JtTeyXI-vv4
part2: http://www.youtube.com/watch?v=6BGIXIqUTLQ