Cooking a dish by 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 cooking Chinese clam dish by following
instructions from a webpage. There are no sound in
the video because I wanted to show the viewers what
the robot is thinking while making the dish. The
flashing text and freeze frames are the internal
thoughts of the robot and not instruction text for
the viewers. These internal thoughts describe the
details of how the robot produce intelligence.
The robot learns to
cook meals by reading instructions in books or
observing demonstrations. This is how human beings
learn to cook. We all know that humans can work in
any restaurant, like: Mcdonalds, Pizza hut, Chinese
restaurants, Mexican restaurants, French restaurants
and so forth. And the majority of knowledge comes
from reading cook books or observing demonstrations.
When reading a
cooking recipe, the robot is using human intelligence
to organize data. He uses logic, common sense, and
organization skills to identify the ingredients,
linear cooking steps, map out important information
from the webpage (this makes searching for data in
the webpage easier), grouping tasks together,
establishing association between task and rules (for
example, establishing relations from cooking steps to
ingredients needed for each step), ignoring
information not related to cooking clams, and
filtering out long or useless data from webpage.
By the way, a webpage
is like a page from a cook book.
After reading the
recipe instructions 1 or 2 times, the robot should
have all the information it needs to cook Chinese
clams. He will know the linear cooking steps (a
general outline of the steps and a detailed outline
of the steps), know the ingredients needed to make
the dish, and know how to organize the instructions
so that he won't run into trouble during the cooking
phase.
Occasionally the
robot will forget things (like what is step4 or what
are the ingredients needed for preparing the sauce),
however, he can re-look at the webpage and find the
answers. Because the instructions to cooking Chinese
clams are so long and the ingredients is long, the
robot will occasionally forget steps. The robot uses
logic to search for forgotten information from the
webpage.
Practice making the
same dish 2-3 times will allow the robot to have
permanent knowledge about cooking Chinese clam dish.
In the second try, the robot still might occasionally
look at instructions in cook books/webpages, but
after the 3rd try, the robot knows how to make
Chinese clam dish using knowledge from its brain. It
can recall all ingredients and linear steps from
memory, instead of looking that information up in
cook books. Remember, the robot's brain stores
pathways in memory. These pathways record
instructions from cook books, as well as, the robot's
5 senses of making the Chinese clam dish.
This method of
learning to cook meals by reading cook books can be
used to learn to cook "any" meal. If the robot
doesn't know how to cook a lobster dish, he can read
the instructions in a book or watch someone do a
demonstration. If the robot doesn't know how to make
pizza, he can read the instructions in a book or
watch someone do a demonstration. If the robot
doesn't know how to make clam chowder, he can read
the instructions in a book or watch someone do a
demonstration. The robot can learn to cook any meal
by reading cook books.
This is very
important because if you want to build a robot to
cook in a French restaurant or a Chinese restaurant,
this type of learning must be used. The current AI
methods used to teach robots to cook are useless.
Things like machine learning or programmers manually
inserting knowledge, or common sense systems are
completely useless. The reason that no one today has
built a robot that can work in a restaurant is strong
proof that current AI methods do not work (2013).
By the way, the
method to learn to cook can be used to learn complex
subject matters, like computer science, or physics,
or chemistry. When reading technical books, the robot
has to organize information, identify important data,
filter bad data, establish relations between diagrams
and text, identify a process, identify linear steps,
generate a rule system, reinterpreting vague
information, summarizing text, and so forth, are used
to learn complex subject matters. In computer
science, a neural network has many parts and has many
processes. The robot has to use intelligence to
understand a neural network from all aspects. Also,
the job of the robot's brain is to organize this data
and store said data in optimal areas in memory so
that the search functions can find information fast.