Semantic networks are a knowledge representation scheme.
This lesson will cover the following topics:
In each knowledge representation, there is a language, and that language has a vocabulary. In addition, the representation contains some content (or knowledge).
$$ F = ma $$ Force is equal to mass times acceleration
How to represent Raven’s Progressive Matrices using a semantic network. State A, and state B.
Lexicon: Consider each node to be a unique state, represented by: - number of prisoners and guards on left side - number of prisoners and guards on right side - side that boat is on.
Which transitions (e.g. moves) between states are both legal AND productive? Represent total possible states given transformations possible:
|trans:||init:<||2p to R:>||p to L:<||2p to R:>||p to L:<||2g to R:>||g.p to L:<||2g to R:>||p to L:<||2p to R:>||p to L:<||2p to R:<|
|\(g_i\)||3, 0||3, 0||3, 0||3, 0||3, 0||1, 2||2, 1||0, 3||0, 3||0, 3||0, 3||0, 3|
|\(p_i\)||3, 0||1, 2||2, 1||0, 3||1, 2||1, 2||2, 1||2, 1||3, 0||1, 2||2, 0||0, 3|
One can weigh transformations to favor specific types of transformations over others. For example:
Takes one state, and generates all possible states, however, should not generate states that are not productive.
Analyzes all generated states, and filters to only legal (i.e., those states that adhere to logical constraints) and productive (i.e., resultant states that have not yet been observed) states.
Note: For each problem, we have to find the right balance the smartness of generators and testers.