Discover
concept correlations routine searches for correlations between concepts and
updates correspondent cause-effect relations.
Discover
concept correlations routine is a part of learning process.
Discover
concept correlations routine assumes that simultaneously activated
concepts are related to each other.
Input
parameter: {cause-effect relation increment}
Every axon
which links two active concepts is updated in
accordance with this formula:
new coherence
attribute value =
= old coherence attribute value +
+ {cause-effect axon increment}
1) Look at short memory.
2) Add {cause-effect
axon increment} value to the coherence attribute value of the correspondent cause-effect relation.
1) Basic plain text parsing routine –
run the discover concept correlations routine after a portion of text is
loaded into short memory.
2) Every thought may end up with discover concept correlations
routine.
3) Any softcoded routine may call discover concept
correlations routine if necessary.
4) Deliberation routine may call discover
concept correlations routine if necessary.
Let’s
imagine that basic plain text
parsing routine reads this text into short memory: “The Sun is a
medium-sized star”. Let {cause-effect
axon increment} = 10.
Then discover
concept correlations routine creates (or updates coherence attribute) these cause-effect relations:
“the” ->
“sun”
“sun” ->
“the”
“sun” ->
“is”
“is” ->
“sun”
.....
“sun” ->
“star”
“star”
-> “sun”
“the” ->
“star”
“star”
-> “the”
Coherence
field of a newly created relation (from this list) will be set to 10.
Coherence
field of already created relation (from this list) will be increased by 10.