Playing Sonic 2 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
playing a videogame. There are no sound in the video
because I wanted to show the viewers what the robot
is thinking while playing the game. 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.
My robot doesn't use:
planning programs/heuristic searches (used by MIT and
Stanford University), Bayesian's probability theories
for decision making, Bayesian's equation for
induction and deduction, semantic networks for
natural language understanding, predicate calculus,
common sense systems, first-order logic, rule-based
systems, genetic programming, or MACHINE LEARNING.
I think having the
ability to predict the future is the most important
aspect of human intelligence. In the video, i show
viewers of an example of how the robot predicts the
future. When playing fast paced games like racing or
sonic the hedgehog, there is no time to think. In
order to play these fast paced games the robot has to
predict the future (general or detailed or both). If
the robot can predict future events he can anticipate
future obstacles. If the robot can predict future
obstacles, he can select options to avoid obstacles.
This fast pace action
game is just a simple example of how the robot
predicts the future. If the robot was to play sonic
the hedgehog, there is no way he can beat the game
without predicting future events. The reason why is
because the scenes in the game move really fast. The
robot will not have time to think and take action
during real time. His brain has to have a set of
future actions that are generated and ready to go
before future events happen.
For example, in the
video, at the ending, the robot was able to predict
future actions before future events happened. The
scene is a roll of snakes lined up on a stream. The
first segment of snakes are wide apart and the player
can jump over these snakes, even with high speed.
The second segment of snakes are compact, and they
require slow speed in order to jump over. The robot
has to understand that and to slow down prior to
encountering segment2 snakes. The scenes in the game
move so fast that if the robot doesn't know there are
compact snakes ahead, he won't be able to slow down
in time. The robot needs to know there are compact
snakes ahead when encountering the first segment of
snakes. This gives the robot time to slow down the
player and to avoid losing a life.
How exactly does the
robot predict the future? It's very hard to predict
the future for unknown games. However, the robot
will be using similar games to predict the future for
unknown games. If the robot has played a game
numerous times, he will predict the future based on
pathways (experiences) in memory. If scenes in a
videogame doesn't change much and they look the same
every time, the robot can extract that future event
and use it to understand what's going to happen in
the future. The robot's brain is actually filtering
out noises and focusing on important objects in the
future pathways. He can extract a general future
event or a detailed future event. Usually the robot
generates both general and detailed future pathways.
After predicting the
future, the robot can make decisions on what type of
actions to take in the future. By the way, the robot
predicts the future every second (like an iterative
loop). He is creating future actions, modifying
future actions, and aborting future actions based on
the current environment.
The robot learned how to
predict the future based on lessons learned from
teachers. Teachers has taught the robot simple and
complex examples of predicting the future. Some
future prediction skills are based on personal
experiences or knowledge from books. The robot takes
this future prediction skill and apply it to playing
videogames.
As the robot practices
more and more, the intelligent pathway in the robot's
brain become more optimal. The pathways will store
the exact actions to take at any given moment, the
obstacles that will happen, the exact logic and rules
to follow at any given moment, and so forth. After
many practices the robot has the knowledge to play a
game optimally, based on beneficial goals. The more
practice he goes through the better he gets.