Lesson 3 - Semantic Networks

Question 1

What is a semantic network?

Answer 1

A semantic network captures both the objects and the relationships of the objects.

Question 2

What are the components of a knowledge representation?

Answer 2

Ex: Semantic Network

1) A lexicon (eg X, Y, Z): Tells us something about the vocabulary of the representation language
2) A structure (eg unchanged, above, inside): Tells us how the words of the vocabulary can be composed together into complex representations.
3) Semantics (eg deleted): Tells us how the representation allows us to draw inferences so that we can reason.

Question 3

What are the 5 characteristics of good representations?

Answer 3

1) Make relationships explicit
2) Expose natural constraints
3) Bring objects and relations together
4) Excludes extraneous details
5) Transparent, concise, complete, fast, and computable

Question 4

What is Represent & Reason?

Answer 4

An abstract class of problem solving methods where the representations are compared and reasoned to identify a match/solution.

Question 5

How would you augment Represent & Reason with weights?

Answer 5

When given 2 potential answers, a weighted scale can be used to help make a decision based on the relationships.