Page 44 - DCAP506_ARTIFICIAL_INTELLIGENCE

Artificial Intelligence

Notes 3. AI is a very complicated scientific problem, so there are huge advantages in locating parts
of the problem that can be separated out and separately attacked.
4. It is quite hard to formalize the facts of general knowledge. Current programs that influence
facts in some of the domains are restricted to special cases and don’t concern the difficulties
that must be conquer to attain very intelligent behavior.

We will converse what facts a person or robot must consider in order to attain a goal by some
approach of action. We will ignore the question of how these facts are displayed, e.g., whether
they are demonstrated by program. We begin with great generality, so there are many problems.
We get successively simpler problems by presuming that the difficulties we have acknowledged
don’t take place until we obtain to a class of problems we believe we can solve.
1. We start by enquiring whether solving the problem needs the cooperation of other people
or overcoming their opposition. If either is true, there are two subcases. In the first subcase,
the other people’s wishes and goals must be considered, and the actions they will take in
specified conditions predicted on the supposition that they will attempt to attain their
goals, which may have to be exposed. The problem is even more hard if bargaining is
concerned, because then the problems and indeterminacies of game theory are pertinent.
Even if bargaining is not concerned, the robot still must “put himself in the position of the
other people with whom he communicates”.

!

Caution Facts such as a person wanting a thing or a person hating another must be illustrated.
The second subcase builds the assumption that the other people can be considered as
machines with recognized input-output behavior. This is frequently a good assumption,
e.g., one considers that a clerk in a store will sell the goods in exchange for their price and
that a professor will allocate a grade in accordance with the quality of the work completed.
Neither the goals of the clerk or the professor need be considered; either might well
regard an effort to access them to optimize the communication as an incursion of privacy.

In such conditions, man generally prefers to be treated as a machine. Let us now presume
that either other people are not concerned in the problem or that the information obtainable
regarding their actions takes the form of input-output relations and does not entail
understanding their goals.

2. The second question is whether the strategy includes the acquisition of knowledge. Even
if we can regard other people as machines, we still may have to reason regarding what
they know. Therefore an airline clerk knows what airplanes fly from here to there and
when, even though he will tell you when asked without your having to motivate him.
One must also take into account information in books and in tables. The latter information
is illustrated by other information.
3. The second subcase of knowledge is according to whether the information obtained can be
simply plugged into a program or whether it enters in a more complex manner. Therefore,
if the robot must telephone someone, its program can just dial the number received, but it
might have to ask a question, “How can I get in touch with Mike?” and reason concerning
how to access the resulting information in combination with other information. The
common distinction may be according to whether new sentences are produced or whether
values are just allocated to variables.

Example: An example valued considering is that a sophisticated air traveler rarely enquires
how he will obtain from the arriving flight to the departing flight at an airport where he must

38 LOVELY PROFESSIONAL UNIVERSITY