Playing who wants to be a millionaire
(pt 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 who wants to be a millionaire. 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.
The IBM Watson back
in 2011 played Jeopardy and beat the best human
players (Ken Jennings and the other guy). This robot
works completely different from the IBM Watson. First
of all, my robot doesn't use machine learning (these
are the right answers and these are the wrong
answers). The Watson uses massive servers to store
and process information because it's a computational
system and not an analytical system. My robot doesn't
need to crunch out all possible answers to a question
(millions of them, ranked in descending order). My
robot contains trained pathways that go "directly" to
the answer. Think of the Watson as a million missiles
fired at a target randomly, and my robot as 1 missile
fired at a target intelligently. Both methods will
destroy the target, but my robot took only 1 shot,
while the Watson took a million shots.
There is a reason why
the human brain can do so much, but the space that it
occupies is so small. The IBM Watson, if it was built
efficiently, wouldn't take up more than 1 server. I
believe that if we build my robot correctly, the
brain of this robot won't take up more than 1 server
(2013). This includes storage of data and processing
intelligence in the robot. Believe it or not, the
human brain can store a lot of data (videos, music
files, static data, linear procedures, and memories)
in a very tiny space. Why can't we do the same thing
with a robot brain?
All information to
answer questions is learned from teachers in school
or through personal experience. For example, the
first question asked in the video was about content
in a movie. The robot has to extract the movie,
Gladiator from memory and search for a particular
scene. Once he finds the correct scene, he uses
analytical methods to identify targeted objects in
the scene. This QandA method was learned from
teachers in school. Teachers teach the robot how to
search for information and how to extract information
from pictures.
Another notable
difference between the IBM Watson and my robot is the
Watson can only play Jeopardy (it can't do anything
else). My robot can do "any" human task. It can play
Monopoly, drive a car, cook, write complex software
programs, fly a plane, do college assignments, etc.
The Watson can do other things only if engineers
modify the internal codes. For example, if you want
the Watson to diagnose patients, software engineers
has to spend 5-10 years designing the algorithm.
By the way, if my
human robot was competing with the IBM Watson, my
robot will probably lose because it thinks and acts
like a human being.
In the video, the
robot is actually doing multiple tasks at the same
time. He is following the rules and objectives of the
game, using strategies to beat the game, and
answering questions. For example, if the robot was
given the same question twice, he will tell the judge
that he has already answered the question. This is
based on common sense. He is aware of common sense
things during the game, which is important. In the
Jeopardy challenge, Watson was unable to understand
and sense what answers were given by other players.
Thus, Watson sometimes gave the same answers as other
players.