associative learning

[ERIK M. ALTMANN]

> My first problem:
> Because of continued creation of new chunks and the Prior Strength
> Equation S*ji = ln(m) - ln(n) states shown later are stronger associated
> with their components as compared to states shown earlier. That effect
> is due to the fact that initially there are only about 75 chunks in
> declarative memory and cannot be found in the data.
> - Is the assumption of the initial number of chunks <100 realistic?
>    (If it were 1000 chunks initially, the unwanted effect would almost
> disappear.)
> - Has anybody encountered a similar problem?
> - Are there solutions to circumvent the effect?

I set *wme-number* to about 1000 at the start of a simulation.

Correct performance in my model depends on retrieving the newest out
of many chunks.  These are created in fairly close succession during
the simulation (every 5-10 seconds), so there's not much spare time
for old chunks to decay.  With too few chunks in the system initially,
chunks created early in the simulation have IAs so high that they
overwhelm the effects of base-level decay and the system fails.

I figure the role of sleep is to keep ~1000 < m < ~10000.

Erik.

---------------------------
Erik M. Altmann
ARCH Lab 2E5
George Mason University
Fairfax, VA  22030
703-993-1326
hfac.gmu.edu/~altmann
---------------------------