Page 98 - DCAP506_ARTIFICIAL_INTELLIGENCE

Artificial Intelligence

Notes 7.3 Non-Monotonic Reasoning

In monotonic reasoning if we expand at set of axioms we cannot withdraw any present
declarations or axioms.

Example: Predicate logic and the conclusions we execute on it is an example of monotonic
reasoning.
Humans do not hold to this monotonic structure when reasoning:
 We are required to jump to conclusions so as to plan and, more fundamentally, survive.

 We cannot expect all potential results of our plan.

!
Caution We must make suppositions regarding things we do not purposely recognize
about.
7.4 Default Reasoning

This is a very general from of non-monotonic reasoning. Here we want to sketch conclusion based
on what is most probable to be correct.

We will converse about two strategies to do this:
 Non-Monotonic logic.
 Default logic.

!
Caution Do not get puzzled regarding the label Non-Monotonic and Default being useful to
reasoning and a specific logic.

Did u know? Non-Monotonic reasoning is common depiction of a class of reasoning.
Non-Monotonic logic is a particular theory. The same is considered for Default reasoning and
Default logic.

Non-Monotonic Logic

This is fundamentally an extension of first-order predicate logic to comprise a modal operator, M.
The reason of this is to permit for steadiness.

Example:

 x: plays_instrument(x) M improvises(x)  jazz_musician(x)
specifies that for all x, x plays an instrument and if the truth that x can manage is reliable with all
other knowledge then we can conclude that x is a jazz musician.

Now we will define consistency:

92 LOVELY PROFESSIONAL UNIVERSITY