Page 97 - DCAP506_ARTIFICIAL_INTELLIGENCE

Unit 7: Symbolic Reasoning under Uncertainty

3. Reasoning by analogy: Humans are good at this, more complicated for AI systems. E.g. If Notes
we are inquired Can robins fly? The system may reason that robins are like sparrows and it
knows sparrows can fly .
4. Generalization and abstraction: Again humans are efficient at this. This is fundamentally
obtaining towards learning and understanding techniques.

5. Meta-level reasoning: Once again utilizes knowledge regarding what you know and
possibly ordering it in some type of significance.

Task Illustrate the concept of reasoning by analogy.

Self Assessment

Fill in the blanks:
1. .............................. reasoning implies to basic rules of inference along with logic knowledge
representations.
2. .............................. reasoning uses procedures that state how to possibly solve (sub) problems.

7.2 Uncertain Reasoning

Regrettably the world is an uncertain place. Any AI system that looks for a model and reasoning
in such a world must be able to deal with this.
Particularly it must be able to contract with:
 Incompleteness — compensate for be deficiency of knowledge.

 Inconsistencies — resolve indistinctness and contradictions.
 Change — it must be able to modernize its world knowledge base over time.

Notes Obviously so as to deal with this, some decision that is made are more possible to
be true (or false) than others and we must bring in techniques that can manage with this
uncertainty.
There are three fundamental methods that can do this:

 Symbolic methods.
 Statistical methods.
 Fuzzy logic methods.

Self Assessment

Fill in the blank:

3. .............................. lead to compensate for deficiency of knowledge.

LOVELY PROFESSIONAL UNIVERSITY 91