Information Technology Reference
In-Depth Information
Fig. 14.8
Sample Output from SNARK
2. There is a low-level
tutti
iff every player has a level greater than zero and less
than five, and no player has a solo.
Tutti
(
low
)
↔∀
p
:
Person
∃
l
:
Level
Level
(
p
,
l
)
∧
LessThan
(
l
,
f
)
∧
LessThan
(
a
,
l
))
∧¬
Solo
(
p
)
3. There is a high-level
tutti
iff every player has a level greater than four.
Tutti
(
high
)
↔∀
p
:
Person
∃
l
:
Level
(
Level
(
p
,
l
)
∧
LessThan
(
e
,
l
))
4. It is the end iff all players are at level zero.
End
↔∀
p
:
Person
(
Level
(
p
,
a
))
By determining states logically in this fashion, the agent is guided to play in a
fashion that is expected on a social level. However, given that the state determines
certain weightings within the FILTER machine-learning aspect, the agent remains
able to improvise freely within the expected range, as a response to the other players;
in effect, this is the same as a human player reasoning about what the other players
are doing and what they intend, and reacting accordingly. Similarly, when the agent
is first initialized, it has no knowledge about what combinations of inputs relate to
Search WWH ::
Custom Search